ocs_ci.ocs package¶

Subpackages¶

Submodules¶

ocs_ci.ocs.amq module¶

AMQ Class to run amq specific tests

class ocs_ci.ocs.amq.AMQ(**kwargs)¶

Bases: object

Workload operation using AMQ

cleanup(kafka_namespace='myproject', tiller_namespace='tiller')¶

Clean up function, will start to delete from amq cluster operator then amq-connector, persistent, bridge, at the end it will delete the created namespace

Parameters

kafka_namespace (str) – Created namespace for amq
tiller_namespace (str) – Created namespace for benchmark

create_consumer_pod(num_of_pods=1, value='10000')¶

Creates producer pods

Parameters

num_of_pods (int) – Number of consumer pods to be created
value (str) – Number of messages to be received

Returns: consumer pod object

create_kafka_topic(name='my-topic', partitions=1, replicas=1)¶

Creates kafka topic

Parameters

name (str) – Name of the kafka topic
partitions (int) – Number of partitions
replicas (int) – Number of replicas

Return: kafka_topic object

create_kafka_user(name='my-user')¶

Creates kafka user

Parameters: name (str) – Name of the kafka user

Return: kafka_user object

create_messaging_on_amq(topic_name='my-topic', user_name='my-user', partitions=1, replicas=1, num_of_producer_pods=1, num_of_consumer_pods=1, value='10000')¶

Creates workload using Open Messaging tool on amq cluster

Parameters

topic_name (str) – Name of the topic to be created
user_name (str) – Name of the user to be created
partitions (int) – Number of partitions of topic
replicas (int) – Number of replicas of topic
num_of_producer_pods (int) – Number of producer pods to be created
num_of_consumer_pods (int) – Number of consumer pods to be created
value (str) – Number of messages to be sent and received

create_namespace(namespace)¶

create namespace for amq

Parameters: namespace (str) – Namespace for amq pods

create_producer_pod(num_of_pods=1, value='10000')¶

Creates producer pods

Parameters

num_of_pods (int) – Number of producer pods to be created
value (str) – Number of the messages to be sent

Returns: producer pod object

export_amq_output_to_gsheet(amq_output, sheet_name, sheet_index)¶

Collect amq data to google spreadsheet

Parameters

amq_output (dict) – amq output in dict
sheet_name (str) – Name of the sheet
sheet_index (int) – Index of sheet

is_amq_pod_running(pod_pattern, expected_pods, namespace='myproject')¶

The function checks if provided pod_pattern finds a pod and if the status is running or not

Parameters

pod_pattern (str) – the pattern for pod
expected_pods (int) – Number of pods
namespace (str) – Namespace for amq pods

Returns

status of pod: True if found pod is running

Return type

bool

run_amq_benchmark(benchmark_pod_name='benchmark', kafka_namespace='myproject', tiller_namespace='tiller', num_of_clients=8, worker=None, timeout=1800, amq_workload_yaml=None, run_in_bg=False)¶

Run benchmark pod and get the results

Parameters

benchmark_pod_name (str) – Name of the benchmark pod
kafka_namespace (str) – Namespace where kafka cluster created
tiller_namespace (str) – Namespace where tiller pod needs to be created
num_of_clients (int) – Number of clients to be created
worker (str) – Loads to create on workloads separated with commas e.g http://benchmark-worker-0.benchmark-worker:8080, http://benchmark-worker-1.benchmark-worker:8080
timeout (int) – Time to complete the run
amq_workload_yaml (dict) – Contains amq workloads information keys and values :name (str): Name of the workloads :topics (int): Number of topics created :partitions_per_topic (int): Number of partitions per topic :message_size (int): Message size :payload_file (str): Load to run on workload :subscriptions_per_topic (int): Number of subscriptions per topic :consumer_per_subscription (int): Number of consumers per subscription :producers_per_topic (int): Number of producers per topic :producer_rate (int): Producer rate :consumer_backlog_sizegb (int): Size of block in gb :test_duration_minutes (int): Time to run the workloads
run_in_bg (bool) – On true the workload will run in background

Returns

Returns benchmark run information if run_in_bg is False.: Otherwise a thread of the amq workload execution

Return type

result (str/Thread obj)

run_amq_workload(command, benchmark_pod_name, tiller_namespace, timeout)¶

Runs amq workload in bg

Parameters

command (str) – Command to run on pod
benchmark_pod_name (str) – Pod name
tiller_namespace (str) – Namespace of pod
timeout (int) – Time to complete the run

Returns

Returns benchmark run information

Return type

result (str)

run_in_bg(namespace='myproject', value='10000', since_time=1800)¶

Validate messages are produced and consumed in bg

Parameters

namespace (str) – Namespace of the pod
value (str) – Number of messages to be sent and received
since_time (int) – Number of seconds to required to sent and receive msg

setup_amq_cluster(sc_name, namespace='myproject', size=100, replicas=3)¶

Creates amq cluster with persistent storage.

Parameters

sc_name (str) – Name of sc
namespace (str) – Namespace for amq cluster
size (int) – Size of the storage
replicas (int) – Number of kafka and zookeeper pods to be created

setup_amq_cluster_operator(namespace='myproject')¶

Function to setup amq-cluster_operator, the file is pulling from github it will make sure cluster-operator pod is running

Parameters: namespace (str) – Namespace for AMQ pods

setup_amq_kafka_bridge()¶

Function to setup amq-kafka, the file file is pulling from github it will make kind: KafkaBridge and will make sure the pod status is running

Return: kafka_bridge object

setup_amq_kafka_connect()¶

The function is to setup amq-kafka-connect, the yaml file is pulling from github it will make kind: KafkaConnect and will make sure the status is running

Returns: kafka_connect object

setup_amq_kafka_persistent(sc_name, size=100, replicas=3)¶

Function to setup amq-kafka-persistent, the file is pulling from github it will make kind: Kafka and will make sure the status is running

Parameters

sc_name (str) – Name of sc
size (int) – Size of the storage in Gi
replicas (int) – Number of kafka and zookeeper pods to be created

return : kafka_persistent

validate_amq_benchmark(result, amq_workload_yaml, benchmark_pod_name='benchmark')¶

Validates amq benchmark run

Parameters

result (str) – Benchmark run information
amq_workload_yaml (dict) – AMQ workload information
benchmark_pod_name (str) – Name of the benchmark pod

Returns

Returns the dict output on success, Otherwise none

Return type

res_dict (dict)

validate_messages_are_consumed(namespace='myproject', value='10000', since_time=1800)¶

Validates if all messages are received in consumer pod

Parameters

namespace (str) – Namespace of the pod
value (str) – Number of messages are recieved
since_time (int) – Number of seconds to required to receive the msg

Raises exception on failures

validate_messages_are_produced(namespace='myproject', value='10000', since_time=1800)¶

Validates if all messages are sent in producer pod

Parameters

namespace (str) – Namespace of the pod
value (str) – Number of messages are sent
since_time (int) – Number of seconds to required to sent the msg

Raises exception on failures

validate_msg(pod, namespace='myproject', value='10000', since_time=1800)¶

Validate if messages are sent or received

Parameters

pod (str) – Name of the pod
namespace (str) – Namespace of the pod
value (str) – Number of messages are sent
since_time (int) – Number of seconds to required to sent the msg

Returns

True if all messages are sent/received

Return type

bool

ocs_ci.ocs.api_client module¶

A module for implementing specific api client for interacting with openshift or kubernetes cluster

APIClientBase is an abstract base class which imposes a contract on methods to be implemented in derived classes which are specific to api client

class ocs_ci.ocs.api_client.APIClientBase¶

Bases: object

Abstract base class for all api-client classes

This is an abstract base class and api-client specific classes should implement all the methods for interacting with openshift cluster

abstract api_create()¶

abstract api_delete()¶

abstract api_get()¶

abstract api_patch()¶

abstract api_post()¶

abstract create_service(**kw)¶

Create an openshift service

Parameters

**kw – ex body=”content”, namespace=’ocnamespace’

Returns

Response from api server

Note: Returns could be tricky because it varies from client to client. If its oc-cli then extra work needs to be done in the specific implementation.

Return type

dict

abstract get_labels(pod_name, pod_namespace)¶

Get the openshift labels on a given pod

Parameters

pod_name (str) – Name of pod for which labels to be fetched
pod_namespace (str) – namespace of pod where this pod lives

Raises

NotImplementedError – if function not implemented by client

Returns

Labels associated with pod

Return type

dict

abstract get_pods(**kwargs)¶

Because of discrepancies in IO format of each client api leaving this to be implemented by specific client

Parameters: **kwargs – ex: namespace=’namespace’,openshift namespace from which we need pods
Raises: NotImplementedError – if client has not implemented this function.
Returns: A list of pod names
Return type: pod_names (list)

abstract property name¶: Concrete class will have respective api-client name

class ocs_ci.ocs.api_client.KubeClient¶

Bases: ocs_ci.ocs.api_client.APIClientBase

All activities using upstream kubernetes python client

property name¶: Concrete class will have respective api-client name

class ocs_ci.ocs.api_client.OCCLIClient¶

Bases: ocs_ci.ocs.api_client.APIClientBase

All activities using oc-cli.

This implements all functionalities like create, patch, delete using oc commands.

property name¶: Concrete class will have respective api-client name

class ocs_ci.ocs.api_client.OCRESTClient¶

Bases: ocs_ci.ocs.api_client.APIClientBase

All activities using openshift REST client

api_create(**kw)¶

api_delete(**kw)¶

api_get(**kw)¶

api_patch(**kw)¶

api_post(**kw)¶

create_service(**kw)¶

Parameters: kw – ex: body={body} for the request which has service spec
Returns: ResourceInstance

get_labels(pod_name, pod_namespace)¶

Get labels from a specific pod

Parameters

pod_name (str) – Name of the pod
pod_namespace (str) – namespace where this pod lives

Raises

NotFoundError – If resource not found

Returns

All the labels on a pod

Return type

dict

get_pods(**kwargs)¶

Get pods in specific namespace or across oc cluster

Parameters

**kwargs – ex: namespace=rook-ceph, label_selector=’x==y’

Returns

of pods names,if no namespace provided then this function: returns all pods across openshift cluster

Return type

list

property name¶: Concrete class will have respective api-client name

ocs_ci.ocs.api_client.get_api_client(client_name)¶

Get instance of corresponding api-client object with given name

Parameters: client_name (str) – name of the api client to be instantiated
Returns: api client object

ocs_ci.ocs.bucket_utils module¶

Helper functions file for working with object buckets

ocs_ci.ocs.bucket_utils.abort_all_multipart_upload(s3_obj, bucketname, object_key)¶

Abort all Multipart Uploads for this Bucket

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier

Returns

List of aborted upload ids

Return type

list

ocs_ci.ocs.bucket_utils.abort_multipart(s3_obj, bucketname, object_key, upload_id)¶

Aborts a Multipart Upload for this Bucket

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier
upload_id (str) – Multipart Upload-ID

Returns

aborted upload id

Return type

str

ocs_ci.ocs.bucket_utils.bucket_read_api(mcg_obj, bucket_name)¶

Fetches the bucket metadata like size, tiers etc

Parameters

mcg_obj (obj) – MCG object
bucket_name (str) – Name of the bucket

Returns

Bucket policy response

Return type

dict

ocs_ci.ocs.bucket_utils.check_pv_backingstore_status(backingstore_name, namespace=None, desired_status=['`OPTIMAL`', '`LOW_CAPACITY`'])¶

check if existing pv backing store is in OPTIMAL state

Parameters

backingstore_name (str) – backingstore name
namespace (str) – backing store’s namespace
desired_status (str) – desired state for the backing store, if None is given then desired
the Healthy status (is) –

Returns

True if backing store is in the desired state

Return type

bool

ocs_ci.ocs.bucket_utils.cli_create_aws_backingstore(mcg_obj, cld_mgr, backingstore_name, uls_name, region)¶

Create a new backingstore with aws underlying storage using noobaa cli command

Parameters

mcg_obj (MCG) – Used for execution for the NooBaa CLI command
cld_mgr (CloudManager) – holds secret for backingstore creation
backingstore_name (str) – backingstore name
uls_name (str) – underlying storage name
region (str) – which region to create backingstore (should be the same as uls)

ocs_ci.ocs.bucket_utils.cli_create_azure_backingstore(mcg_obj, cld_mgr, backingstore_name, uls_name, region)¶

Create a new backingstore with aws underlying storage using noobaa cli command

Parameters

cld_mgr (CloudManager) – holds secret for backingstore creation
backingstore_name (str) – backingstore name
uls_name (str) – underlying storage name
region (str) – which region to create backingstore (should be the same as uls)

ocs_ci.ocs.bucket_utils.cli_create_google_backingstore(mcg_obj, cld_mgr, backingstore_name, uls_name, region)¶

Create a new backingstore with GCP underlying storage using a NooBaa CLI command

Parameters

mcg_obj (MCG) – Used for execution for the NooBaa CLI command
cld_mgr (CloudManager) – holds secret for backingstore creation
backingstore_name (str) – backingstore name
uls_name (str) – underlying storage name
region (str) – which region to create backingstore (should be the same as uls)

ocs_ci.ocs.bucket_utils.cli_create_ibmcos_backingstore(mcg_obj, cld_mgr, backingstore_name, uls_name, region)¶

Create a new backingstore with IBM COS underlying storage using a NooBaa CLI command

Parameters

cld_mgr (CloudManager) – holds secret for backingstore creation
backingstore_name (str) – backingstore name
uls_name (str) – underlying storage name
region (str) – which region to create backingstore (should be the same as uls)

ocs_ci.ocs.bucket_utils.cli_create_pv_backingstore(mcg_obj, backingstore_name, vol_num, size, storage_class)¶

Create a new backingstore with pv underlying storage using noobaa cli command

Parameters

backingstore_name (str) – backingstore name
vol_num (int) – number of pv volumes
size (int) – each volume size in GB
storage_class (str) – which storage class to use

ocs_ci.ocs.bucket_utils.cli_create_s3comp_backingstore(cld_mgr, backingstore_name, uls_name, region)¶

ocs_ci.ocs.bucket_utils.compare_directory(awscli_pod, original_dir, result_dir, amount=2)¶

Compares object checksums on original and result directories

Args:
awscli_pod (pod): A pod running the AWS CLI tools original_dir (str): original directory name result_dir (str): result directory name amount (int): Number of test objects to create

ocs_ci.ocs.bucket_utils.complete_multipart_upload(s3_obj, bucketname, object_key, upload_id, parts)¶

Completes the Multipart Upload

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier
upload_id (str) – Multipart Upload-ID
parts (list) – List containing the uploaded parts which includes ETag and part number

Returns

Dictionary containing the completed multipart upload details

Return type

dict

ocs_ci.ocs.bucket_utils.craft_s3_command(cmd, mcg_obj=None, api=False, signed_request_creds=None)¶

Crafts the AWS CLI S3 command including the login credentials and command to be ran

Parameters

mcg_obj – An MCG object containing the MCG S3 connection credentials
cmd – The AWSCLI command to run
api – True if the call is for s3api, false if s3
signed_request_creds – a dictionary containing AWS S3 creds for a signed request

Returns

The crafted command, ready to be executed on the pod

Return type

str

ocs_ci.ocs.bucket_utils.create_multipart_upload(s3_obj, bucketname, object_key)¶

Initiates Multipart Upload

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket on which multipart upload to be initiated on
object_key (str) – Unique object Identifier

Returns

Multipart Upload-ID

Return type

str

ocs_ci.ocs.bucket_utils.del_objects(uploaded_objects_paths, awscli_pod, mcg_obj)¶

Deleting objects from bucket

Parameters

uploaded_objects_paths (list) – List of object paths
awscli_pod (pod) – A pod running the AWSCLI tools
mcg_obj (obj) – An MCG object containing the MCG S3 connection credentials

ocs_ci.ocs.bucket_utils.delete_bucket_policy(s3_obj, bucketname)¶

Deletes bucket policy

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket

Returns

Delete Bucket policy response

Return type

dict

ocs_ci.ocs.bucket_utils.get_bucket_available_size(mcg_obj, bucket_name)¶

Function to get the bucket available size

Parameters

mcg_obj (obj) – MCG object
bucket_name (str) – Name of the bucket

Returns

Available size in the bucket

Return type

int

ocs_ci.ocs.bucket_utils.get_bucket_policy(s3_obj, bucketname)¶

Gets bucket policy from a bucket

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket

Returns

Get Bucket policy response

Return type

dict

ocs_ci.ocs.bucket_utils.get_full_path_object(downloaded_files, bucket_name)¶

Getting full of object in the bucket

Parameters

downloaded_files (list) – List of downloaded files
bucket_name (str) – Name of the bucket

Returns

List of full paths of objects

Return type

uploaded_objects_paths (list)

ocs_ci.ocs.bucket_utils.get_rgw_restart_counts()¶

Gets the restart count of the RGW pods

Returns: restart counts of RGW pods
Return type: list

ocs_ci.ocs.bucket_utils.list_multipart_upload(s3_obj, bucketname)¶

Lists the multipart upload details on a bucket

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket

Returns

Dictionary containing the multipart upload details

Return type

dict

ocs_ci.ocs.bucket_utils.list_uploaded_parts(s3_obj, bucketname, object_key, upload_id)¶

Lists uploaded parts and their ETags

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier
upload_id (str) – Multipart Upload-ID

Returns

Dictionary containing the multipart upload details

Return type

dict

ocs_ci.ocs.bucket_utils.namespace_bucket_update(mcg_obj, bucket_name, read_resource, write_resource)¶

Edits MCG namespace bucket resources

Parameters

mcg_obj (obj) – An MCG object containing the MCG S3 connection credentials
bucket_name (str) – Name of the bucket
read_resource (list) – Resource names to provide read access
write_resource (str) – Resource name to provide write access

ocs_ci.ocs.bucket_utils.obc_io_create_delete(mcg_obj, awscli_pod, bucket_factory)¶: Running IOs on OBC interface :param mcg_obj: An MCG object containing the MCG S3 connection credentials :type mcg_obj: obj :param awscli_pod: A pod running the AWSCLI tools :type awscli_pod: pod :param bucket_factory: Calling this fixture creates a new bucket(s)

ocs_ci.ocs.bucket_utils.oc_create_aws_backingstore(cld_mgr, backingstore_name, uls_name, region)¶

Create a new backingstore with aws underlying storage using oc create command

Parameters

cld_mgr (CloudManager) – holds secret for backingstore creation
backingstore_name (str) – backingstore name
uls_name (str) – underlying storage name
region (str) – which region to create backingstore (should be the same as uls)

ocs_ci.ocs.bucket_utils.oc_create_azure_backingstore(cld_mgr, backingstore_name, uls_name, region)¶

Create a new backingstore with Azure underlying storage using oc create command

Parameters

cld_mgr (CloudManager) – holds secret for backingstore creation
backingstore_name (str) – backingstore name
uls_name (str) – underlying storage name
region (str) – which region to create backingstore (should be the same as uls)

ocs_ci.ocs.bucket_utils.oc_create_google_backingstore(cld_mgr, backingstore_name, uls_name, region)¶

Create a new backingstore with GCP underlying storage using oc create command

Parameters

cld_mgr (CloudManager) – holds secret for backingstore creation
backingstore_name (str) – backingstore name
uls_name (str) – underlying storage name
region (str) – which region to create backingstore (should be the same as uls)

ocs_ci.ocs.bucket_utils.oc_create_ibmcos_backingstore(cld_mgr, backingstore_name, uls_name, region)¶

Create a new backingstore with IBM COS underlying storage using oc create command

Parameters

cld_mgr (CloudManager) – holds secret for backingstore creation
backingstore_name (str) – backingstore name
uls_name (str) – underlying storage name
region (str) – which region to create backingstore (should be the same as uls)

ocs_ci.ocs.bucket_utils.oc_create_pv_backingstore(backingstore_name, vol_num, size, storage_class)¶

Create a new backingstore with pv underlying storage using oc create command

Parameters

backingstore_name (str) – backingstore name
vol_num (int) – number of pv volumes
size (int) – each volume size in GB
storage_class (str) – which storage class to use

ocs_ci.ocs.bucket_utils.oc_create_s3comp_backingstore(cld_mgr, backingstore_name, uls_name, region)¶

ocs_ci.ocs.bucket_utils.put_bucket_policy(s3_obj, bucketname, policy)¶

Adds bucket policy to a bucket

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
policy (json) – Bucket policy in Json format

Returns

Bucket policy response

Return type

dict

ocs_ci.ocs.bucket_utils.retrieve_anon_s3_resource()¶

Returns an anonymous boto3 S3 resource by creating one and disabling signing

Disabling signing isn’t documented anywhere, and this solution is based on a comment by an AWS developer: https://github.com/boto/boto3/issues/134#issuecomment-116766812

Returns: An anonymous S3 resource
Return type: boto3.resource()

ocs_ci.ocs.bucket_utils.retrieve_test_objects_to_pod(podobj, target_dir)¶

Downloads all the test objects to a given directory in a given pod.

Parameters

podobj (OCS) – The pod object to download the objects to
target_dir – The fully qualified path of the download target folder

Returns

A list of the downloaded objects’ names

Return type

list

ocs_ci.ocs.bucket_utils.retrieve_verification_mode()¶

ocs_ci.ocs.bucket_utils.rm_object_recursive(podobj, target, mcg_obj, option='')¶

Remove bucket objects with –recursive option

Parameters

podobj (OCS) – The pod on which to execute the commands and download the objects to
target (str) – Fully qualified bucket target path
mcg_obj (MCG, optional) – The MCG object to use in case the target or source are in an MCG
option (str) – Extra s3 remove command option

ocs_ci.ocs.bucket_utils.s3_copy_object(s3_obj, bucketname, source, object_key)¶

Boto3 client based copy object

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
source (str) – Source object key. eg: ‘<bucket>/<key>
object_key (str) – Unique object Identifier for copied object

Returns

Copy object response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_delete_bucket_website(s3_obj, bucketname)¶

Boto3 client based Delete bucket website function

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket

Returns

DeleteBucketWebsite response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_delete_object(s3_obj, bucketname, object_key, versionid=None)¶

Simple Boto3 client based Delete object

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier
versionid (str) – Unique version number of an object

Returns

Delete object response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_delete_objects(s3_obj, bucketname, object_keys)¶

Boto3 client based delete objects

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_keys (list) – The objects to delete. Format: {‘Key’: ‘object_key’, ‘VersionId’: ‘’}

Returns

delete objects response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_get_bucket_versioning(s3_obj, bucketname, s3_client=None)¶

Boto3 client based Get Bucket Versioning function

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
s3_client – Any s3 client resource

Returns

GetBucketVersioning response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_get_bucket_website(s3_obj, bucketname)¶

Boto3 client based Get bucket website function

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket

Returns

GetBucketWebsite response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_get_object(s3_obj, bucketname, object_key, versionid='')¶

Simple Boto3 client based Get object

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier
versionid (str) – Unique version number of an object

Returns

Get object response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_get_object_acl(s3_obj, bucketname, object_key)¶

Boto3 client based get_object_acl operation

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier for copied object

Returns

get object acl response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_head_object(s3_obj, bucketname, object_key)¶

Boto3 client based head_object operation to retrieve only metadata

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier for copied object

Returns

head object response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_io_create_delete(mcg_obj, awscli_pod, bucket_factory)¶: Running IOs on s3 bucket :param mcg_obj: An MCG object containing the MCG S3 connection credentials :type mcg_obj: obj :param awscli_pod: A pod running the AWSCLI tools :type awscli_pod: pod :param bucket_factory: Calling this fixture creates a new bucket(s)

ocs_ci.ocs.bucket_utils.s3_list_object_versions(s3_obj, bucketname, prefix='')¶

Boto3 client based list object Versionfunction

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
prefix (str) – Object key prefix

Returns

List object version response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_list_objects_v1(s3_obj, bucketname, prefix='', delimiter='', max_keys=1000, marker='')¶

Boto3 client based list object version1

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
prefix (str) – Limits the response to keys that begin with the specified prefix.
delimiter (str) – Character used to group keys.
max_keys (int) – Maximum number of keys returned in the response. Default 1,000 keys.
marker (str) – key to start with when listing objects in a bucket.

Returns

list object v1 response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_list_objects_v2(s3_obj, bucketname, prefix='', delimiter='', max_keys=1000, con_token='', fetch_owner=False)¶

Boto3 client based list object version2

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
prefix (str) – Limits the response to keys that begin with the specified prefix.
delimiter (str) – Character used to group keys.
max_keys (int) – Maximum number of keys returned in the response. Default 1,000 keys.
con_token (str) – Token used to continue the list
fetch_owner (bool) – Unique object Identifier

Returns

list object v2 response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_put_bucket_versioning(s3_obj, bucketname, status='Enabled', s3_client=None)¶

Boto3 client based Put Bucket Versioning function

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
status (str) – ‘Enabled’ or ‘Suspended’. Default ‘Enabled’
s3_client – Any s3 client resource

Returns

PutBucketVersioning response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_put_bucket_website(s3_obj, bucketname, website_config)¶

Boto3 client based Put bucket website function

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
website_config (dict) – Website configuration info

Returns

PutBucketWebsite response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_put_object(s3_obj, bucketname, object_key, data, content_type='')¶

Simple Boto3 client based Put object

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
object_key (str) – Unique object Identifier
data (str) – string content to write to a new S3 object
content_type (str) – Type of object data. eg: html, txt etc,

Returns

Put object response

Return type

dict

ocs_ci.ocs.bucket_utils.s3_upload_part_copy(s3_obj, bucketname, copy_source, object_key, part_number, upload_id)¶

Boto3 client based upload_part_copy operation

Parameters

s3_obj (obj) – MCG or OBC object
bucketname (str) – Name of the bucket
copy_source (str) – Name of the source bucket and key name. {bucket}/{key}
part_number (int) – Part number
upload_id (str) – Upload Id
object_key (str) – Unique object Identifier for copied object

Returns

upload_part_copy response

Return type

dict

ocs_ci.ocs.bucket_utils.setup_base_objects(awscli_pod, original_dir, result_dir, amount=2)¶

Prepares two directories and populate one of them with objects

Args:
awscli_pod (Pod): A pod running the AWS CLI tools original_dir (str): original directory name result_dir (str): result directory name amount (Int): Number of test objects to create

ocs_ci.ocs.bucket_utils.sync_object_directory(podobj, src, target, s3_obj=None, signed_request_creds=None)¶

Syncs objects between a target and source directories

Parameters

podobj (OCS) – The pod on which to execute the commands and download the objects to
src (str) – Fully qualified object source path
target (str) – Fully qualified object target path
s3_obj (MCG, optional) – The MCG object to use in case the target or source are in an MCG
signed_request_creds (dictionary, optional) – the access_key, secret_key, endpoint and region to use when willing to send signed aws s3 requests

ocs_ci.ocs.bucket_utils.upload_parts(mcg_obj, awscli_pod, bucketname, object_key, body_path, upload_id, uploaded_parts)¶

Uploads individual parts to a bucket

Parameters

mcg_obj (obj) – An MCG object containing the MCG S3 connection credentials
awscli_pod (pod) – A pod running the AWSCLI tools
bucketname (str) – Name of the bucket to upload parts on
object_key (list) – Unique object Identifier
body_path (str) – Path of the directory on the aws pod which contains the parts to be uploaded
upload_id (str) – Multipart Upload-ID
uploaded_parts (list) – list containing the name of the parts to be uploaded

Returns

List containing the ETag of the parts

Return type

list

ocs_ci.ocs.bucket_utils.verify_s3_object_integrity(original_object_path, result_object_path, awscli_pod)¶

Verifies checksum between original object and result object on an awscli pod

Parameters

original_object_path (str) – The Object that is uploaded to the s3 bucket
result_object_path (str) – The Object that is downloaded from the s3 bucket
awscli_pod (pod) – A pod running the AWSCLI tools

Returns

True if checksum matches, False otherwise

Return type

bool

ocs_ci.ocs.bucket_utils.wait_for_pv_backingstore(backingstore_name, namespace=None)¶

wait for existing pv backing store to reach OPTIMAL state

Parameters

backingstore_name (str) – backingstore name
namespace (str) – backing store’s namespace

ocs_ci.ocs.bucket_utils.write_individual_s3_objects(mcg_obj, awscli_pod, bucket_factory, downloaded_files, target_dir, bucket_name=None)¶

Writes objects one by one to an s3 bucket

Parameters

mcg_obj (obj) – An MCG object containing the MCG S3 connection credentials
awscli_pod (pod) – A pod running the AWSCLI tools
bucket_factory – Calling this fixture creates a new bucket(s)
downloaded_files (list) – List of downloaded object keys
target_dir (str) – The fully qualified path of the download target folder
bucket_name (str) – Name of the bucket (default: none)

ocs_ci.ocs.clients module¶

class ocs_ci.ocs.clients.WinNode(**kw)¶

Bases: object

check_disk(number)¶

connect_to_target(ip, username, password)¶

create_disk(number)¶

create_fio_job_options(job_options)¶

create_new_target(ip, port=3260)¶

delete_target()¶

disconnect_from_target()¶

get_iscsi_initiator_name()¶

run_fio_test()¶

start_iscsi_initiator()¶

win_exec(ps_command, timeout=180)¶

ocs_ci.ocs.cluster module¶

A module for all rook functionalities and abstractions.

This module has rook related classes, support for functionalities to work with rook cluster. This works with assumptions that an OCP cluster is already functional and proper configurations are made for interaction.

class ocs_ci.ocs.cluster.CephCluster¶

Bases: object

Handles all cluster related operations from ceph perspective

This class has depiction of ceph cluster. Contains references to pod objects which represents ceph cluster entities.

pods¶

A list of ceph cluster related pods

Type: list

cluster_name¶

Name of ceph cluster

Type: str

namespace¶

openshift Namespace where this cluster lives

Type: str

calc_trim_mean_throughput(samples=8)¶

Calculate the cluster average throughput out of a few samples

Parameters: samples (int) – The number of samples to take
Returns: The average cluster throughput
Return type: float

check_ceph_pool_used_space(cbp_name)¶

Check for the used space of a pool in cluster

Returns:
used_in_gb (float): Amount of used space in pool (in GBs)

Raises:
UnexpectedBehaviour: If used size keeps varying in Ceph status

cluster_health_check(timeout=None)¶

Check overall cluster health. Relying on health reported by CephCluster.get()

Parameters: timeout (int) – in seconds. By default timeout value will be scaled based on number of ceph pods in the cluster. This is just a crude number. Its been observed that as the number of pods increases it takes more time for cluster’s HEALTH_OK.
Returns: True if “HEALTH_OK” else False
Return type: bool
Raises: CephHealthException – if cluster is not healthy

property cluster_name

create_user(username, caps)¶

Create a ceph user in the cluster

Parameters

username (str) – ex client.user1
caps (str) – ceph caps ex: mon ‘allow r’ osd ‘allow rw’

Returns

return value of get_user_key()

get_admin_key()¶

Returns: base64 encoded key
Return type: adminkey (str)

get_ceph_capacity()¶

The function gets the total mount of storage capacity of the ocs cluster. the calculation is <Num of OSD> * <OSD size> / <replica number> it will not take into account the current used capacity.

Returns: Total storage capacity in GiB (GiB is for development environment)
Return type: int

get_ceph_cluster_iops()¶

The function gets the IOPS from the ocs cluster

Returns: Total IOPS in the cluster

get_ceph_health(detail=False)¶

Exec ceph health cmd on tools pod and return the status of the ceph cluster.

Parameters: detail (bool) – If True the ‘ceph health detail’ is executed
Returns: Output of the ceph health command.
Return type: str

get_ceph_status(format=None)¶

Exec ceph status cmd on tools pod and return its output.

Parameters: format (str) – Format of the output (e.g. json-pretty, json, plain)
Returns: Output of the ceph status command.
Return type: str

get_cluster_throughput()¶

Function to get the throughput of ocs cluster

Returns: The write throughput of the cluster in MiB/s
Return type: float

get_iops_percentage(osd_size=2)¶

The function calculates the IOPS percentage of the cluster depending on number of osds in the cluster

Parameters: osd_size (int) – Size of 1 OSD in Ti
Returns: IOPS percentage of the OCS cluster

get_mons_from_cluster()¶

Getting the list of mons from the cluster

Returns: Returns the mons from the cluster
Return type: available_mon (list)

get_rebalance_status()¶

This function gets the rebalance status

Returns: True if rebalance is completed, False otherwise
Return type: bool

get_throughput_percentage()¶

Function to get throughput percentage of the ocs cluster

Returns: Throughput percentage of the cluster

get_user_key(user)¶

Parameters: user (str) – ceph username ex: client.user1
Returns: base64 encoded user key
Return type: key (str)

is_health_ok()¶

Returns: True if “HEALTH_OK” else False
Return type: bool

property mcg_obj¶

mds_change_count(new_count)¶

Change mds count in the cluster

Parameters: new_count (int) – Absolute number of active mdss required

mds_health_check(count)¶

MDS health check based on pod count

Parameters: count (int) – number of pods expected
Raises: MDACountException – if pod count doesn’t match

mon_change_count(new_count)¶

Change mon count in the cluster

Parameters: new_count (int) – Absolute number of mons required

mon_health_check(count)¶

Mon health check based on pod count

Parameters: count (int) – Expected number of mon pods
Raises: MonCountException – if mon pod count doesn’t match

property namespace

noobaa_health_check()¶: Check Noobaa health

property pods

remove_mon_from_cluster()¶

Removing the mon pod from deployment

Returns: True if removal of mon is successful, False otherwise
Return type: remove_mon(bool)

scan_cluster()¶: Get accurate info on current state of pods

set_noout()¶: Set noout flag for maintainance

static set_port(pod)¶

Set port attribute on pod. port attribute for mon is required for secrets and this attrib is not a member for original pod class.

Parameters: pod (Pod) – Pod object without ‘port’ attribute
Returns: A modified pod object with ‘port’ attribute set
Return type: pod(Pod)

time_taken_to_complete_rebalance(timeout=600)¶

This function calculates the time taken to complete rebalance

Parameters: timeout (int) – Time to wait for the completion of rebalance
Returns: Time taken in minutes for the completion of rebalance
Return type: int

unset_noout()¶: unset noout flag for peering

wait_for_noobaa_health_ok(tries=60, delay=5)¶: Wait for Noobaa health to be OK

wait_for_rebalance(timeout=600)¶

Wait for re-balance to complete

Parameters: timeout (int) – Time to wait for the completion of re-balance
Returns: True if rebalance completed, False otherwise
Return type: bool

class ocs_ci.ocs.cluster.CephClusterExternal¶

Bases: ocs_ci.ocs.cluster.CephCluster

Handle all external ceph cluster related functionalities Assumption: Cephcluster Kind resource exists

cluster_health_check(timeout=300)¶: This would be a comprehensive cluster health check which includes checking pods, external ceph cluster health. raise exceptions.CephHealthException(“Cluster health is NOT OK”)

property cluster_name¶

property namespace¶

validate_pvc()¶: Check whether all PVCs are in bound state

wait_for_cluster_cr()¶: we have to wait for cluster cr to appear else it leads to list index out of range error

wait_for_nooba_cr()¶

class ocs_ci.ocs.cluster.CephHealthMonitor(ceph_cluster, sleep=5)¶

Bases: threading.Thread

Context manager class for monitoring ceph health status of CephCluster. If CephCluster will get to HEALTH_ERROR state it will save the ceph status to health_error_status variable and will stop monitoring.

log_error_status()¶

run()¶

Method representing the thread’s activity.

You may override this method in a subclass. The standard run() method invokes the callable object passed to the object’s constructor as the target argument, if any, with sequential and keyword arguments taken from the args and kwargs arguments, respectively.

ocs_ci.ocs.cluster.calculate_compression_ratio(pool_name)¶

Calculating the compression of data on RBD pool

Parameters: pool_name (str) – the name of the pool to calculate the ratio on
Returns: the compression ratio in percentage
Return type: int

ocs_ci.ocs.cluster.check_ceph_osd_tree()¶

Checks whether an OSD tree is created/modified correctly. It is a summary of the previous functions: ‘check_osd_tree_1az_vmware’, ‘check_osd_tree_3az_cloud’, ‘check_osd_tree_1az_cloud’.

Returns: True, if the ceph osd tree is formed correctly. Else False
Return type: bool

ocs_ci.ocs.cluster.check_ceph_osd_tree_after_node_replacement()¶

Check the ceph osd tree after the process of node replacement.

Returns: True if the ceph osd tree formation is correct, and all the OSD’s are up. Else False
Return type: bool

ocs_ci.ocs.cluster.check_osd_tree_1az_cloud(osd_tree, number_of_osds)¶

Checks whether an OSD tree is created/modified correctly. This can be used as a verification step for deployment and cluster expansion tests. This function is specifically for ocs cluster created on 1 AZ config

Parameters

osd_tree (dict) – Dictionary of the values which represent ‘osd tree’.
number_of_osds (int) – total number of osds in the cluster

Returns

True, if the ceph osd tree is formed correctly. Else False

Return type

Boolean

ocs_ci.ocs.cluster.check_osd_tree_1az_vmware(osd_tree, number_of_osds)¶

Checks whether an OSD tree is created/modified correctly. This can be used as a verification step for deployment and cluster expansion tests. This function is specifically for ocs cluster created on 1 AZ VMWare setup

Parameters

osd_tree (dict) – Dictionary of the values which represent ‘osd tree’.
number_of_osds (int) – total number of osds in the cluster

Returns

True, if the ceph osd tree is formed correctly. Else False

Return type

bool

ocs_ci.ocs.cluster.check_osd_tree_3az_cloud(osd_tree, number_of_osds)¶

Checks whether an OSD tree is created/modified correctly. This can be used as a verification step for deployment and cluster expansion tests. This function is specifically for ocs cluster created on 3 AZ config

Parameters

osd_tree (dict) – Dictionary of the values which represent ‘osd tree’.
number_of_osds (int) – total number of osds in the cluster

Returns

True, if the ceph osd tree is formed correctly. Else False

Return type

Boolean

ocs_ci.ocs.cluster.check_osds_in_hosts_are_up(osd_tree)¶

Check if all the OSD’s in status ‘up’

Parameters: osd_tree (dict) – The ceph osd tree
Returns: True if all the OSD’s in status ‘up’. Else False
Return type: bool

ocs_ci.ocs.cluster.check_osds_in_hosts_osd_tree(hosts, osd_tree)¶

Checks if osds are formed correctly after cluster expansion

Parameters

hosts (list) – List of hosts
osd_tree (str) – ‘ceph osd tree’ command output

Returns

True if osd tree formatted correctly

Return type

bool

ocs_ci.ocs.cluster.count_cluster_osd()¶

The function returns the number of cluster OSDs

Returns: number of OSD pods in current cluster
Return type: osd_count (int)

ocs_ci.ocs.cluster.get_balancer_eval()¶

Function to get ceph pg balancer eval value

Returns: Eval output of pg balancer
Return type: eval_out (float)

ocs_ci.ocs.cluster.get_byte_used_by_pool(pool_name)¶

Check byte_used value for specific pool

Parameters: pool_name (str) – name of the pool to check replica
Returns: The amount of byte stored from pool.
Return type: integer

ocs_ci.ocs.cluster.get_ceph_df_detail()¶

Get ceph osd df detail

Returns: ‘ceph df details’ command output
Return type: dict

ocs_ci.ocs.cluster.get_child_nodes_osd_tree(node_id, osd_tree)¶

This function finds the children of a node from the ‘ceph osd tree’ and returns them as list

Parameters

node_id (int) – the id of the node for which the children to be retrieved
osd_tree (dict) – dictionary containing the output of ‘ceph osd tree’

Returns

of ‘children’ of a given node_id

Return type

list

ocs_ci.ocs.cluster.get_osd_pods_memory_sum()¶

Get the sum of memory of all OSD pods. This is used to determine the size needed for a PVC so when IO will be running over it the OSDs cache will be filled

Returns: The sum of the OSD pods memory in GB
Return type: int

ocs_ci.ocs.cluster.get_osd_utilization()¶

Get osd utilization value

Returns: Dict of osd name and its used value i.e {‘osd.1’: 15.276289408185841, ‘osd.0’: 15.276289408185841, ‘osd.2’: 15.276289408185841}
Return type: osd_filled (dict)

ocs_ci.ocs.cluster.get_percent_used_capacity()¶

Function to calculate the percentage of used capacity in a cluster

Returns: The percentage of the used capacity in the cluster
Return type: float

ocs_ci.ocs.cluster.get_pg_balancer_status()¶

Function to check pg_balancer active and mode is upmap

Returns: True if active and upmap is set else False
Return type: bool

ocs_ci.ocs.cluster.get_pgs_per_osd()¶

Function to get ceph pg count per OSD

Returns: Dict of osd name and its used value i.e {‘osd.0’: 136, ‘osd.2’: 136, ‘osd.1’: 136}
Return type: osd_dict (dict)

ocs_ci.ocs.cluster.is_lso_cluster()¶

Check if the cluster is an lso cluster

Returns: True, if the cluster is an lso cluster. False, otherwise
Return type: bool

ocs_ci.ocs.cluster.silence_ceph_osd_crash_warning(osd_pod_name)¶

Silence the osd crash warning of a specific osd pod

Parameters: osd_pod_name (str) – The name of the osd pod which we need to silence the crash warning
Returns: True if it found the osd crash with name ‘osd_pod_name’. False otherwise
Return type: bool

ocs_ci.ocs.cluster.validate_cluster_on_pvc()¶

Validate creation of PVCs for MON and OSD pods. Also validate that those PVCs are attached to the OCS pods

Raises: AssertionError – If PVC is not mounted on one or more OCS pods

ocs_ci.ocs.cluster.validate_compression(pool_name)¶

Check if data was compressed

Parameters: pool_name (str) – name of the pool to check replica
Returns: True if compression works. False if not
Return type: bool

ocs_ci.ocs.cluster.validate_ocs_pods_on_pvc(pods, pvc_names)¶

Validate if ocs pod has PVC. This validation checking if there is the pvc like: rook-ceph-mon-a for the pod rook-ceph-mon-a-56f67f5968-6j4px.

Parameters

pods (list) – OCS pod names
pvc_names (list) – names of all PVCs

Raises

AssertionError – If no PVC found for one of the pod

ocs_ci.ocs.cluster.validate_osd_utilization(osd_used=80)¶

Validates osd utilization matches osd_used value

Parameters

osd_used (int) – osd used value

Returns

True if all osd values is equal or greater to osd_used.: False Otherwise.

Return type

bool

ocs_ci.ocs.cluster.validate_pdb_creation()¶

Validate creation of PDBs for MON, MDS and OSD pods.

Raises: AssertionError – If required PDBs were not created.

ocs_ci.ocs.cluster.validate_pg_balancer()¶

Validate either data is equally distributed to OSDs

Returns: True if avg PG’s per osd difference is <=10 else False
Return type: bool

ocs_ci.ocs.cluster.validate_replica_data(pool_name, replica)¶

Check if data is replica 2 or 3

Parameters

replica (int) – size of the replica(2,3)
pool_name (str) – name of the pool to check replica

Returns

True if replicated data size is meet rep config and False if dont

Return type

Bool

ocs_ci.ocs.cluster.wait_for_silence_ceph_osd_crash_warning(osd_pod_name, timeout=900)¶

Wait for ‘timeout’ seconds to check for the ceph osd crash warning, and silence it.

Parameters

osd_pod_name (str) – The name of the osd pod which we need to silence the crash warning
timeout (int) – time in seconds to wait for silence the osd crash warning

Returns

True if it found the osd crash with name ‘osd_pod_name’. False otherwise

Return type

bool

ocs_ci.ocs.cluster_load module¶

A module for cluster load related functionalities

class ocs_ci.ocs.cluster_load.ClusterLoad(project_factory=None, pvc_factory=None, sa_factory=None, pod_factory=None, target_percentage=None)¶

Bases: object

A class for cluster load functionalities

adjust_load_if_needed()¶: Dynamically adjust the IO load based on the cluster latency. In case the latency goes beyond 250 ms, start deleting FIO pods. Once latency drops back below 100 ms, re-create the FIO pods to make sure that cluster load is around the target percentage

calc_trim_metric_mean(metric, samples=5, mute_logs=False)¶

Get the trimmed mean of a given metric

Parameters

metric (str) – The metric to calculate the average result for
samples (int) – The number of samples to take
mute_logs (bool) – True for muting the logs, False otherwise

Returns

The average result for the metric

Return type

float

decrease_load(wait=True)¶

Delete DeploymentConfig with its pods and the PVC. Then, wait for the IO to be stopped

Parameters: wait (bool) – True for waiting for IO to drop after the deletion of the FIO pod, False otherwise

get_query(query, mute_logs=False)¶

Get query from Prometheus and parse it

Parameters

query (str) – Query to be done
mute_logs (bool) – True for muting the logs, False otherwise

Returns

the query result

Return type

float

increase_load(rate, wait=True)¶

Create a PVC, a service account and a DeploymentConfig of FIO pod

Parameters

rate (str) – FIO ‘rate’ value (e.g. ‘20M’)
wait (bool) – True for waiting for IO to kick in on the newly created pod, False otherwise

increase_load_and_print_data(rate, wait=True)¶

Increase load and print data

Parameters

rate (str) – FIO ‘rate’ value (e.g. ‘20M’)
wait (bool) – True for waiting for IO to kick in on the newly created pod, False otherwise

print_metrics(mute_logs=False)¶

Print metrics

Parameters: mute_logs (bool) – True for muting the Prometheus logs, False otherwise

reach_cluster_load_percentage()¶: Reach the cluster limit and then drop to the given target percentage. The number of pods needed for the desired target percentage is determined by creating pods one by one, while examining the cluster latency. Once the latency is greater than 250 ms and it is growing exponentially, it means that the cluster limit has been reached. Then, dropping to the target percentage by deleting all pods and re-creating ones with smaller value of FIO ‘rate’ param. This leaves the number of pods needed running IO for cluster load to be around the desired percentage.

reduce_load(pause=True)¶: Pause the cluster load

resume_load()¶: Resume the cluster load

ocs_ci.ocs.cluster_load.wrap_msg(msg)¶

Wrap a log message with ‘=’ marks. Necessary for making cluster load background logging distinguishable

Parameters: msg (str) – The log message to wrap
Returns: The wrapped log message
Return type: str

ocs_ci.ocs.constants module¶

Constants module.

This module contains any values that are widely used across the framework, utilities, or tests that will predominantly remain unchanged.

In the event values here have to be changed it should be under careful review and with consideration of the entire project.

ocs_ci.ocs.couchbase module¶

Couchbase workload class

class ocs_ci.ocs.couchbase.CouchBase(**kwargs)¶

Bases: ocs_ci.ocs.pillowfight.PillowFight

CouchBase workload operation

WAIT_FOR_TIME = 1800¶

admission_parts = ['/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/workloads/couchbase/server/admissionServiceAccount.yaml', '/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/workloads/couchbase/server/admissionClusterRole.yaml', '/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/workloads/couchbase/server/admissionClusterRoleBinding.yaml', '/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/workloads/couchbase/server/admissionSecret.yaml', '/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/workloads/couchbase/server/admissionDeployment.yaml', '/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/workloads/couchbase/server/admissionService.yaml', '/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/workloads/couchbase/server/MutatingWebhookConfiguration.yaml', '/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/workloads/couchbase/server/ValidatingWebhookConfiguration.yaml']¶

admission_pod = []¶

analyze_run(skip_analyze=False)¶

Analyzing the workload run logs

Parameters: skip_analyze (bool) – Option to skip logs analysis

cb_examples = <ocs_ci.ocs.resources.ocs.OCS object>¶

cb_worker = <ocs_ci.ocs.resources.ocs.OCS object>¶

couchbase_pod = <ocs_ci.ocs.ocp.OCP object>¶

create_couchbase_worker(replicas=1, sc_name=None)¶

Deploy a Couchbase server and pillowfight workload using operator

The couchbase workers do not come up unless there is an admission controller running. The admission controller is started from the default project prior to bringing up the operator. Secrets, rolebindings and serviceaccounts need to also be generated.

Once the couchbase operator is running, we need to wait for the three worker pods to also be up. Then a pillowfight task is started.

After the pillowfight task has finished, the log is collected and analyzed.

Raises: Exception – If pillowfight results indicate that a minimum performance level is not reached (1 second response time, less than 1000 ops per second)

get_couchbase_nodes()¶

Get nodes that contain a couchbase app pod

Returns: List of nodes
Return type: list

is_up_and_running(pod_name, ocp_value)¶

Test if the pod specified is up and running.

Parameters

pod_name (str) – Name of pod being checked.
ocp_value (object) – object used for running oc commands

Returns

bool; True if pod is running, False otherwise

ns_obj = <ocs_ci.ocs.ocp.OCP object>¶

pod_obj = <ocs_ci.ocs.ocp.OCP object>¶

respin_couchbase_app_pod()¶

Respin the couchbase app pod

Returns: pod status

run_workload(replicas, num_items=None, num_threads=None, run_in_bg=False)¶: Running workload with pillow fight operator :param replicas: Number of pods :type replicas: int :param num_items: Number of items to be loaded to the cluster :type num_items: int :param num_threads: Number of threads :type num_threads: int :param run_in_bg: Optional run IOs in background :type run_in_bg: bool

secretsadder = <ocs_ci.ocs.ocp.OCP object>¶

setup_cb()¶: Creating admission parts,couchbase operator pod, couchbase worker secret

teardown()¶: Delete objects created in roughly reverse order of how they were created.

ocs_ci.ocs.defaults module¶

Defaults module. All the defaults used by OSCCI framework should reside in this module. PYTEST_DONT_REWRITE - avoid pytest to rewrite, keep this msg here please!

ocs_ci.ocs.disruptive_operations module¶

ocs_ci.ocs.disruptive_operations.get_ocs_operator_node_name()¶

Getting node’s name that running ocs-operator pod

Returns: node’s name that running ocs-operator pod
Return type: str

ocs_ci.ocs.disruptive_operations.osd_node_reboot()¶

Rebooting worker node that running OSD

Raises: AssertionError – in case the ceph-tools pod was not recovered

ocs_ci.ocs.disruptive_operations.worker_node_shutdown(abrupt)¶

Shutdown worker node that running ocs-operator pod

Parameters: abrupt – (bool): True if abrupt shutdown, False for permanent shutdown
Raises: AssertionError – in case the ceph-tools pod was not recovered

ocs_ci.ocs.elasticsearch module¶

Deploying an Elasticsearch server for collecting logs from ripsaw benchmarks. Interface for the Performance ElasticSearch server

class ocs_ci.ocs.elasticsearch.ElasticSearch¶

Bases: object

ElasticSearch Environment

cleanup()¶: Cleanup the environment from all Elasticsearch components, and from the port forwarding process.

dumping_all_data(target_path)¶

Dump All data from the internal ES server to .tgz file.

Parameters

target_path (str) – the path where the results file will be copy into

Returns

True if the dump operation succeed and return the results data to the host: otherwise False

Return type

bool

get_health()¶

This method return the health status of the Elasticsearch.

Returns: True if the status is green (OK) otherwise - False
Return type: bool

get_indices()¶

Getting list of all indices in the ES server - all created by the test, the installation of the ES was without any indexes pre-installed.

Returns: list of all indices defined in the ES server
Return type: list

get_ip()¶

This function return the IP address of the Elasticsearch cluster. this IP is to use inside the OCP cluster

Return: str : String that represent the Ip Address.

get_password()¶

This method return the password used to connect the Elasticsearch.

Returns: The password as text
Return type: str

get_port()¶

This function return the port of the Elasticsearch cluster.

Return: str : String that represent the port.

ocs_ci.ocs.elasticsearch.elasticsearch_load(connection, target_path)¶

Load all data from target_path/results into an elasticsearch (es) server.

Parameters

connection (obj) – an elasticsearch connection object
target_path (str) – the path where data was dumped into

Returns

True if loading data succeed, False otherwise

Return type

bool

ocs_ci.ocs.exceptions module¶

exception ocs_ci.ocs.exceptions.CPUNotSufficientException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.CSVNotFound¶: Bases: Exception

exception ocs_ci.ocs.exceptions.CephHealthException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ChannelNotFound¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ClassCreationException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.CommandFailed¶: Bases: Exception

exception ocs_ci.ocs.exceptions.CredReqSecretNotFound¶: Bases: Exception

exception ocs_ci.ocs.exceptions.DeploymentPlatformNotSupported¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ExternalClusterDetailsException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.FailedToAddNodeException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.FailedToDeleteInstance¶: Bases: Exception

exception ocs_ci.ocs.exceptions.FailedToRemoveNodeException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.FipsNotInstalledException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.FlexyDataNotFound¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ImageIsNotDeletedOrNotFound¶: Bases: Exception

exception ocs_ci.ocs.exceptions.InvalidStatusCode¶: Bases: Exception

exception ocs_ci.ocs.exceptions.MDSCountException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.MemoryNotSufficientException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.MissingRequiredConfigKeyError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.MonCountException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.NoBucketPolicyResponse¶: Bases: Exception

exception ocs_ci.ocs.exceptions.NoInstallPlanForApproveFoundException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.NodeNotFoundError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.NonUpgradedImagesFoundError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.NoobaaConditionException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.NoobaaHealthException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.NotFoundError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.NotSupportedFunctionError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.OpenshiftConsoleSuiteNotDefined¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PSIVolumeCreationFailed¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PSIVolumeDeletionFailed¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PSIVolumeNotInExpectedState¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PendingCSRException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PerformanceException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PoolDataNotErased¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PoolNotCompressedAsExpected¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PoolNotFound¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PoolNotReplicatedAsNeeded¶: Bases: Exception

exception ocs_ci.ocs.exceptions.PvcNotDeleted¶: Bases: Exception

exception ocs_ci.ocs.exceptions.RDMDiskNotFound¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ResourceInUnexpectedState¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ResourceLeftoversException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ResourceNameNotSpecifiedException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ResourceNotFoundError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ResourceWrongStatusException(resource_name, describe_out)¶: Bases: Exception

exception ocs_ci.ocs.exceptions.RhcosImageNotFound¶: Bases: Exception

exception ocs_ci.ocs.exceptions.SameNamePrefixClusterAlreadyExistsException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.ServiceUnavailable¶: Bases: Exception

exception ocs_ci.ocs.exceptions.StorageNotSufficientException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.TagNotFoundException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.TimeoutException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.TimeoutExpiredError(*value)¶

Bases: Exception

message = 'Timed Out'¶

exception ocs_ci.ocs.exceptions.UnavailableBuildException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnavailableResourceException¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnexpectedBehaviour¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnexpectedImage¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnexpectedVolumeType¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnhealthyBucket¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnsupportedBrowser¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnsupportedFeatureError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnsupportedOSType¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnsupportedPlatformError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.UnsupportedPlatformVersionError¶: Bases: Exception

exception ocs_ci.ocs.exceptions.VMMaxDisksReachedException¶: Bases: Exception

ocs_ci.ocs.external_ceph module¶

class ocs_ci.ocs.external_ceph.Ceph(name='ceph', node_list=None)¶

Bases: object

property ceph_demon_stat¶

Retrieves expected numbers for demons of each role

Returns: Ceph demon stats
Return type: dict

get_ceph_demons(role=None)¶

Get Ceph demons list

Returns: list of CephDemon
Return type: list

get_ceph_object(role, order_id=0)¶

Returns single ceph object. If order id is provided returns that occurrence from results list, otherwise returns first occurrence

Parameters

role (str) – Ceph object’s role
order_id (int) – order number of the ceph object

Returns

ceph object

Return type

CephObject

get_ceph_objects(role=None)¶

Get Ceph Object by role. Returns all objects if role is not defined. Ceph object can be Ceph demon, client, installer or generic entity. Pool role is never assigned to Ceph object and means that node has no Ceph objects

Parameters: role (str) – Ceph object’s role as str
Returns: ceph objects
Return type: list

get_metadata_list(role, client=None)¶

Returns metadata for demons of specified role

Parameters

role (str) – ceph demon role
client (CephObject) – Client with keyring and ceph-common

Returns

metadata as json object representation

Return type

list

get_node_by_hostname(hostname)¶

Returns Ceph node by it’s hostname

Parameters: hostname (str) – hostname
Returns: ceph node object
Return type: CephNode

get_nodes(role=None, ignore=None)¶

Get node(s) by role. Return all nodes if role is not defined

Parameters

role (str, RolesContainer) – node’s role. Takes precedence over ignore
ignore (str, RolesContainer) – node’s role to ignore from the list

Returns

nodes

Return type

list

get_osd_by_id(osd_id, client=None)¶

Parameters

osd_id (int) – osd id
client (CephObject) – Client with keyring and ceph-common:

Returns

daemon object or None in case the osd daemon list is: empty

Return type

CephDemon

get_osd_container_name_by_id(osd_id, client=None)¶

Parameters

osd_id (int) – osd id
client (CephObject) – Client with keyring and ceph-common:

get_osd_data_partition(osd_id, client=None)¶

Returns data partition by given osd id

Parameters

osd_id (int) – osd id
client (CephObject) – client, optional

Returns

data path

Return type

str

get_osd_data_partition_path(osd_id, client=None)¶

Returns data partition path by given osd id

Parameters

osd_id (int) – osd id
client (CephObject) – Client with keyring and ceph-common:

Returns

data partition path

Return type

str

get_osd_device(osd_id, client=None)¶

Parameters

osd_id (int) – osd id
client (CephObject) – Client with keyring and ceph-common:

Returns

osd device

Return type

str

get_osd_metadata(osd_id, client=None)¶

Retruns metadata for osd by given id

Parameters

osd_id (int) – osd id
client (CephObject) – Client with keyring and ceph-common

Returns

osd metadata like:

{
   "id": 8,
   "arch": "x86_64",
   "back_addr": "172.16.115.29:6801/1672",
   "back_iface": "eth0",
   "backend_filestore_dev_node": "vdd",
   "backend_filestore_partition_path": "/dev/vdd1",
   "ceph_version": "ceph version 12.2.5-42.el7cp (82d52d7efa6edec70f6a0fc306f40b89265535fb) luminous
           (stable)",
   "cpu": "Intel(R) Xeon(R) CPU E5-2690 v3 @ 2.60GHz",
   "osd_data": "/var/lib/ceph/osd/ceph-8",
   "osd_journal": "/var/lib/ceph/osd/ceph-8/journal",
   "osd_objectstore": "filestore",
   "rotational": "1"
}

Return type

dict

get_osd_service_name(osd_id, client=None)¶

Parameters

osd_id (int) – osd id
client (CephObject) – Client with keyring and ceph-common:

Returns

service name

Return type

str

role_to_node_mapping()¶: Given a role we should be able to get the corresponding CephNode object

class ocs_ci.ocs.external_ceph.CephClient(role, node)¶: Bases: ocs_ci.ocs.external_ceph.CephObject

class ocs_ci.ocs.external_ceph.CephDemon(role, node)¶

Bases: ocs_ci.ocs.external_ceph.CephObject

ceph_demon_by_container_name(container_name)¶

property container_name¶

property container_prefix¶

exec_command(cmd, **kw)¶

Proxy to node’s exec_command with wrapper to run commands inside the container for containerized demons

Parameters

cmd (str) – command to execute
**kw – options

Returns

node’s exec_command resut

class ocs_ci.ocs.external_ceph.CephNode(**kw)¶

Bases: object

connect(timeout=300)¶: connect to ceph instance using paramiko ssh protocol eg: self.connect() - setup tcp keepalive to max retries for active connection - set up hostname and shortname as attributes for tests to query

create_ceph_object(role)¶

Create ceph object on the node

Parameters: role (str) – ceph object role
Returns: created ceph object
Return type: CephObject|CephDemon

exec_command(**kw)¶

execute a command on the vm eg: self.exec_cmd(cmd=’uptime’) or self.exec_cmd(cmd=’background_cmd’, check_ec=False)

check_ec¶

False will run the command and not wait for exit code

Type: bool

get_allocated_volumes()¶

get_ceph_demons(role=None)¶

Get Ceph demons list. Only active (those which will be part of the cluster) demons are shown.

Returns: list of CephDemon
Return type: list

get_ceph_objects(role=None)¶

Get Ceph objects list on the node

Parameters: role (str) – Ceph object role
Returns: ceph objects
Return type: list

get_free_volumes()¶

obtain_root_permissions(path)¶

Transfer ownership of root to current user for the path given. Recursive.

Parameters: path (str) – file path

reconnect()¶

remove_ceph_object(ceph_object)¶

Removes ceph object form the node

Parameters: ceph_object (CephObject) – ceph object to remove

property role¶

search_ethernet_interface(ceph_node_list)¶

Search interface on the given node node which allows every node in the cluster accesible by it’s shortname.

Parameters: ceph_node_list (list) – list of CephNode
Returns: returns None if no suitable interface found
Return type: eth_interface (str)

set_eth_interface(eth_interface)¶: set the eth interface

set_internal_ip()¶: set the internal ip of the vm which differs from floating ip

write_file(**kw)¶

class ocs_ci.ocs.external_ceph.CephObject(role, node)¶

Bases: object

exec_command(cmd, **kw)¶

Proxy to node’s exec_command

Parameters

cmd (str) – command to execute
**kw – options

Returns

node’s exec_command result

property pkg_type¶

write_file(**kw)¶

Proxy to node’s write file

Parameters: **kw – options
Returns: node’s write_file result

class ocs_ci.ocs.external_ceph.CephObjectFactory(node)¶

Bases: object

CLIENT_ROLES = ['client']¶

DEMON_ROLES = ['mon', 'osd', 'mgr', 'rgw', 'mds', 'nfs']¶

create_ceph_object(role)¶

Create an appropriate Ceph object by role

Parameters: role – role string
Returns: Ceph object based on role

class ocs_ci.ocs.external_ceph.CephOsd(node, device=None)¶

Bases: ocs_ci.ocs.external_ceph.CephDemon

property container_name¶

property is_active¶

class ocs_ci.ocs.external_ceph.NodeVolume(status)¶

Bases: object

ALLOCATED = 'allocated'¶

FREE = 'free'¶

class ocs_ci.ocs.external_ceph.RolesContainer(role='pool')¶

Bases: object

Container for single or multiple node roles. Can be used as iterable or with equality ‘==’ operator to check if role is present for the node. Note that ‘==’ operator will behave the same way as ‘in’ operator i.e. check that value is present in the role list.

append(object)¶

clear()¶

equals(other)¶

extend(iterable)¶

remove(object)¶

update_role(roles_list)¶

class ocs_ci.ocs.external_ceph.SSHConnectionManager(ip_address, username, password, look_for_keys=False, outage_timeout=300)¶

Bases: object

property client¶

get_client()¶

get_transport()¶

property transport¶

ocs_ci.ocs.fio_artefacts module¶

ocs_ci.ocs.fio_artefacts.get_configmap_dict()¶

ConfigMap template for fio workloads. Note that you need to add actual configuration to workload.fio file.

Returns: YAML data for a OCP ConfigMap object
Return type: dict

ocs_ci.ocs.fio_artefacts.get_job_dict()¶

Job template for fio workloads.

Returns: YAML data for a job object
Return type: dict

ocs_ci.ocs.fio_artefacts.get_mcg_conf(mcg_obj, workload_bucket, custom_options=None)¶

Basic fio configuration for upgrade utilization for NooBaa S3 bucket.

Parameters

mcg_obj (obj) – MCG object, it can be found among fixtures
workload_bucket (obj) – MCG bucket
custom_options (dict) – Dictionary of lists containing tuples with additional configuration for fio in format: {‘section’: [(‘option’, ‘value’),…],…} e.g. {‘global’:[(‘name’,’bucketname’)],’create’:[(‘time_based’,’1’),(‘runtime’,’48h’)]} Those values can be added to the config or rewrite already existing values

Returns

updated fio configuration

Return type

str

ocs_ci.ocs.fio_artefacts.get_pvc_dict()¶

PVC template for fio workloads. Note that all ‘None’ values needs to be defined before usage.

Returns: YAML data for a PVC object
Return type: dict

ocs_ci.ocs.fiojob module¶

This module contains functions which implements functionality necessary to run general fio workloads as k8s Jobs in OCP/OCS cluster via workload fixtures (see ocs_ci.utility.workloadfixture).

ocs_ci.ocs.fiojob.delete_fio_data(fio_job_file, delete_check_func)¶: Delete fio data by removing the fio job resource, with a wait to make sure date were reclaimed on the ceph level.

ocs_ci.ocs.fiojob.fio_to_dict(fio_output)¶: ” Parse fio output and provide parsed dict it as a result.

ocs_ci.ocs.fiojob.get_ceph_storage_stats(ceph_pool_name)¶

Get ceph storage utilization values from ceph df: total STORED value and MAX AVAIL of given ceph pool, which are important for understanding how much space is already consumed and how much is still available.

Parameters: ceph_pool_name (str) – name of ceph pool where you want to write data
Returns: int: sum of all ceph pool STORED values (Bytes) int: value of MAX AVAIL value of given ceph pool (Bytes)
Return type: tuple

ocs_ci.ocs.fiojob.get_storageutilization_size(target_percentage, ceph_pool_name)¶

For the purpose of the workload storage utilization fixtures, get expected pvc_size based on STORED and MAX AVAIL values (as reported by ceph df) for given ceph pool and target utilization percentage.

This is only approximate, and it won’t work eg. if each pool has different configuration of replication.

Parameters

target_percentage (float) – target total utilization, eg. 0.5 for 50%
ceph_pool_name (str) – name of ceph pool where you want to write data

Returns

pvc_size for storage utilization job (in GiB, rounded)

Return type

int

ocs_ci.ocs.fiojob.get_timeout(fio_min_mbps, pvc_size)¶

Compute how long we will let the job running while writing data to the volume.

Parameters

fio_min_mbps (int) – minimal write speed in MiB/s
pvc_size (int) – size of PVC in GiB, which will be used to writing

Returns

write_timeout in seconds

Return type

int

ocs_ci.ocs.fiojob.wait_for_job_completion(namespace, timeout, error_msg)¶

This is a WORKAROUND of particular ocsci design choices: I just wait for one pod in the namespace, and then ask for the pod again to get it’s name (but it would be much better to just wait for the job to finish instead, then ask for a name of the successful pod and use it to get logs …)

Returns: name of Pod resource of the finished job
Return type: str

ocs_ci.ocs.fiojob.workload_fio_storageutilization(fixture_name, project, fio_pvc_dict, fio_job_dict, fio_configmap_dict, measurement_dir, tmp_path, target_percentage=None, target_size=None, with_checksum=False, keep_fio_data=False, minimal_time=480)¶

This function implements core functionality of fio storage utilization workload fixtures. This is necessary because we can’t parametrize single general fixture over multiple parameters (it would mess with test case id and polarion test case tracking).

It works as a workload fixture, as understood by ocs_ci.utility.workloadfixture module.

When target_percentage is specified, the goal of the fixture is to fill whatever is left so that total cluster utilization reaches the target percentage. This means that in this mode, number of data written depends on both total capacity and current utilization. If the current storage utilization already exceeds the target, the test is skipped.

On the other hand with target_size, you can specify the size of data written by fio directly.

Parameters

fixture_name (str) – name of the fixture using this function (for logging and k8s object labeling purposes)
project (ocs_ci.ocs.ocp.OCP) – OCP object of project in which the Job is deployed, as created by project_factory or project fixture
fio_pvc_dict (dict) – PVC k8s struct for fio target volume
fio_job_dict (dict) – Job k8s struct for fio job
fio_configmap_dict (dict) – configmap k8s struct with fio config file
measurement_dir (str) – reference to a fixture which represents a directory where measurement results are stored, see also ocs_ci.utility.workloadfixture.measure_operation()
tmp_path (pathlib.PosixPath) – reference to pytest tmp_path fixture
target_percentage (float) – target utilization as percentage wrt all usable OCS space, eg. 0.50 means a request to reach 50% of total OCS storage utilization (wrt usable space)
target_size (int) – target size of the PVC for fio to use, eg. 10 means a request for fio to write 10GiB of data
with_checksum (bool) – if true, sha1 checksum of the data written by fio is stored on the volume, and reclaim policy of the volume is changed to Retain so that the volume is not removed during test teardown for later verification runs
keep_fio_data (bool) – If true, keep the fio data after the fio storage utilization is completed. Else if false, deletes the fio data.
minimal_time (int) – Minimal number of seconds to monitor a system. (See more details in the function ‘measure_operation’)

Returns

measurement results with timestamps and other medatada from: ocs_ci.utility.workloadfixture.measure_operation()

Return type

dict

ocs_ci.ocs.fiojob.write_data_via_fio(fio_job_file, write_timeout, pvc_size, target_percentage)¶: Write data via fio Job (specified in tf tmp file) to reach desired utilization level, and keep this level for minimal_time seconds.

ocs_ci.ocs.flowtest module¶

class ocs_ci.ocs.flowtest.BackgroundOps¶

Bases: object

handler(func, *args, **kwargs)¶

Wraps the function to run specific iterations

Returns: True if function runs successfully
Return type: bool

wait_for_bg_operations(bg_ops, timeout=1200)¶

Waits for threads to be completed

Parameters

bg_ops (list) – Futures
timeout (int) – Time in seconds to wait

class ocs_ci.ocs.flowtest.FlowOperations¶

Bases: object

Flow based operations class

add_capacity_entry_criteria()¶

Entry criteria verification function for add capacity operation

Returns: containing the params used in add capacity exit operation
Return type: tuple

add_capacity_exit_criteria(restart_count_before, osd_pods_before)¶

Exit criteria function for Add capacity operation

Parameters

restart_count_before (dict) – Restart counts of pods
osd_pods_before (list) – List of OSD pods before

node_operations_entry_criteria(node_type, number_of_nodes, operation_name='Node Operation', network_fail_time=None)¶

Entry criteria function for node related operations

Parameters

node_type (str) – Type of node
number_of_nodes (int) – Number of nodes
operation_name (str) – Name of the node operation
network_fail_time (int) – Total time to fail the network in a node

Returns

containing the params used in Node operations

Return type

tuple

validate_cluster(cluster_check=False, node_status=False, pod_status=False, operation_name='')¶

Validates various ceph and ocs cluster checks

Parameters

node_status (bool) – Verifies node is Ready
pod_status (bool) – Verifies StorageCluster pods in expected state
operation_name (str) – Name of the operation, to Tag

ocs_ci.ocs.hsbench module¶

class ocs_ci.ocs.hsbench.HsBench¶

Bases: object

Hotsauce S3 benchmark

cleanup()¶: Clear all objects in the associated bucket Clean up deployment config, pvc, pod and test user

create_resource_hsbench()¶

Create resource for hsbench mark test:: Create service account Create PVC Create golang pod

create_test_user()¶: Create a radosgw test user for S3 access

delete_test_user()¶: Delete RGW test user and bucket belong to test user

install_hsbench(timeout=2400)¶: Install HotSauce S3 benchmark: https://github.com/markhpc/hsbench

run_benchmark(num_obj=None, run_mode=None, timeout=None)¶

Running Hotsauce S3 benchmark Usage detail can be found at: https://github.com/markhpc/hsbench

Parameters

num_obj (int) – Maximum number of objects
run_mode (string) – mode types
timeout (int) – timeout in seconds

validate_reshard_process()¶

Validate reshard process

Raises: CommandFailed – If reshard process fails

validate_s3_objects()¶: Validate S3 objects created by hsbench on single bucket

ocs_ci.ocs.jenkins module¶

Jenkins Class to run jenkins specific tests

class ocs_ci.ocs.jenkins.Jenkins(num_of_projects=1, num_of_builds=1)¶

Bases: object

Workload operation using Jenkins

Parameters

projects (iterable) – project names
num_of_builds (int) – number of builds per project

cleanup()¶: Clean up

create_app_jenkins()¶: create application jenkins

create_jenkins_build_config()¶: create jenkins build config

create_jenkins_pvc()¶

create jenkins pvc

Returns: pvc_objs
Return type: List

create_ocs_jenkins_template()¶: Create OCS Jenkins Template

create_project_names()¶: Create project names

export_builds_results_to_googlesheet(sheet_name='E2E Workloads', sheet_index=3)¶

Collect builds results, output to google spreadsheet

Parameters

sheet_name (str) – Name of the sheet
sheet_index (int) – Index of sheet

get_build_duration_time(namespace, build_name)¶

get build duration time

Parameters

namespace (str) – get build in namespace
build_name (str) – the name of the jenkins build

get_build_name_by_pattern(pattern='client', namespace=None, filter=None)¶

Get build name by pattern

Returns: build name
Return type: list

get_builds_obj(namespace)¶

Get all jenkins builds

Returns: jenkins deploy pod objects list
Return type: List

get_builds_sorted_by_number(project)¶

Get builds per project and sort builds by build name number

Parameters: project (str) – project name
Returns: List of Jenkins build OCS obj
Return type: List

get_jenkins_deploy_pods(namespace)¶

Get all jenkins deploy pods

Parameters: namespace (str) – get pods in namespace
Returns: jenkins deploy pod objects list
Return type: pod_objs (list)

get_node_name_where_jenkins_pod_not_hosted(node_type='worker', num_of_nodes=1)¶

get nodes

Parameters

node_type (str) – The node type (e.g. worker, master)
num_of_nodes (int) – The number of nodes to be returned

Returns

List of compute node names

Return type

list

property number_builds_per_project¶

property number_projects¶

print_completed_builds_results()¶: Get builds logs and print them on table

start_build()¶: Start build on jenkins

wait_for_build_to_complete(timeout=900)¶

Wait for build status to reach complete state

Parameters: timeout (int) – Time in seconds to wait

wait_for_jenkins_deploy_status(status, timeout=600)¶

Wait for jenkins deploy pods status to reach running/completed

Parameters

status (str) – status to reach Running or Completed
timeout (int) – Time in seconds to wait

ocs_ci.ocs.machine module¶

ocs_ci.ocs.machine.add_annotation_to_machine(annotation, machine_name)¶: Add annotation to the machine :param annotation: Annotation to be set on the machine :type annotation: str :param eg: annotation = “machine.openshift.io/exclude-node-draining=’’” :param machine_name: machine name :type machine_name: str

ocs_ci.ocs.machine.add_node(machine_set, count)¶

Add new node to the cluster

Parameters

machine_set (str) – Name of a machine set to get increase replica count
count (int) – Count to increase

Returns

True if commands executes successfully

Return type

bool

ocs_ci.ocs.machine.check_machineset_exists(machine_set)¶

Function to check machineset exists or not

Parameters: machine_set (str) – Name of the machine set
Returns: True if machineset exists, else false
Return type: bool

ocs_ci.ocs.machine.create_custom_machineset(role='app', instance_type='m4.xlarge', labels=None, taints=None, zone='a')¶

Function to create custom machineset works only for AWS i.e. Using this user can create nodes with different instance type and role. https://docs.openshift.com/container-platform/4.1/machine_management/creating-machineset.html

Parameters

role (str) – Role type to be added for node eg: it will be app,worker
instance_type (str) – Type of aws instance
label (list) – List of Labels (key, val) to be added to the node
taints (list) – List of taints to be applied
zone (str) – Machineset zone for node creation.

Returns

Created machineset name

Return type

machineset (str)

Raise:: ResourceNotFoundError: Incase machineset creation failed UnsupportedPlatformError: Incase of wrong platform

ocs_ci.ocs.machine.create_ocs_infra_nodes(num_nodes)¶

Create infra node instances

Parameters: num_nodes (int) – Number of instances to be created
Returns: list of instance names
Return type: list

ocs_ci.ocs.machine.delete_custom_machineset(machine_set)¶

Function to delete custom machineset

Parameters

machine_set (str) – Name of the machine set to be deleted
WARN – Make sure it’s not OCS worker node machines set, if so then OCS worker nodes and machine set will be deleted.

Raise:: UnexpectedBehaviour: Incase machineset not deleted

ocs_ci.ocs.machine.delete_machine(machine_name)¶

Deletes a machine

Parameters: machine_name (str) – Name of the machine you want to delete
Raises: CommandFailed – In case yaml_file and resource_name wasn’t provided

ocs_ci.ocs.machine.delete_machine_and_check_state_of_new_spinned_machine(machine_name)¶

Deletes a machine and checks the state of the newly spinned machine

Parameters: machine_name (str) – Name of the machine you want to delete
Returns: True in case of success, False otherwise
Return type: bool

ocs_ci.ocs.machine.get_labeled_nodes(label)¶

Fetches all nodes with specific label.

Parameters: label (str) – node label to look for
Returns: List of names of labeled nodes
Return type: list

ocs_ci.ocs.machine.get_machine_from_machineset(machine_set)¶

Get the machine name from its associated machineset

Parameters: machine_set (str) – Name of the machine set
Returns: Machine names
Return type: List

ocs_ci.ocs.machine.get_machine_from_node_name(node_name)¶

Get the associated machine name for the given node name

Parameters: node_name (str) – Name of the node
Returns: Machine name
Return type: str

ocs_ci.ocs.machine.get_machine_objs(machine_names=None)¶

Get machine objects by machine names

Parameters

machine_names (list) – The machine names to get their objects
None, will return all cluster machines (If) –

Returns

Cluster machine OCS objects

Return type

list

ocs_ci.ocs.machine.get_machine_type(machine_name)¶

Get the machine type (e.g. worker, master)

Parameters: machine_name (str) – Name of the machine
Returns: Type of the machine
Return type: str

ocs_ci.ocs.machine.get_machines(machine_type='worker')¶

Get cluster’s machines according to the machine type (e.g. worker, master)

Parameters: machine_type (str) – The machine type (e.g. worker, master)
Returns: The nodes OCP instances
Return type: list

ocs_ci.ocs.machine.get_machineset_from_machine_name(machine_name)¶

Get the machineset associated with the machine name

Parameters: machine_name (str) – Name of the machine
Returns: Machineset name
Return type: str

ocs_ci.ocs.machine.get_machineset_objs(machineset_names=None)¶

Get machineset objects by machineset names

Parameters

machineset_names (list) – The machineset names to get their objects
None, will return all cluster machines (If) –

Returns

Cluster machineset OCS objects

Return type

list

ocs_ci.ocs.machine.get_machinesets()¶

Get machine sets

Returns: list of machine sets
Return type: machine_sets (list)

ocs_ci.ocs.machine.get_ready_replica_count(machine_set)¶

Get replica count which are in ready state in a machine set

Parameters: machine_set (str) – Machineset name
Returns: replica count which are in ready state
Return type: ready_replica (int)

ocs_ci.ocs.machine.get_replica_count(machine_set)¶

Get replica count of a machine set

Parameters: machine_set (str) – Name of a machine set to get replica count
Returns: replica count of a machine set
Return type: replica count (int)

ocs_ci.ocs.machine.get_storage_cluster(namespace='openshift-storage')¶

Get storage cluster name

Parameters: namespace (str) – Namespace of the resource
Returns: Storage cluster name
Return type: str

ocs_ci.ocs.machine.wait_for_new_node_to_be_ready(machine_set, timeout=600)¶

Wait for the new node to reach ready state

Parameters: machine_set (str) – Name of the machine set
Raises: ResourceWrongStatusException – In case the new spun machine fails to reach Ready state or replica count didn’t match

ocs_ci.ocs.mcg_workload module¶

ocs_ci.ocs.mcg_workload.create_workload_job(job_name, bucket, project, mcg_obj, resource_path, custom_options=None)¶

Creates kubernetes job that should utilize MCG bucket.

Parameters

job_name (str) – Name of the job
bucket (objt) – MCG bucket with S3 interface
project (obj) – OCP object representing OCP project which will be used for the job
mcg_obj (obj) – instance of MCG class
resource_path (str) – path to directory where should be created resources
custom_options (dict) – Dictionary of lists containing tuples with additional configuration for fio in format: {‘section’: [(‘option’, ‘value’),…],…} e.g. {‘global’:[(‘name’,’bucketname’)],’create’:[(‘time_based’,’1’),(‘runtime’,’48h’)]} Those values can be added to the config or rewrite already existing values

Returns

Job object

Return type

obj

ocs_ci.ocs.mcg_workload.get_configmap_dict(fio_job, mcg_obj, bucket, custom_options=None)¶

Fio configmap dictionary with configuration set for MCG workload.

Parameters

fio_job (dict) – Definition of fio job
mcg_obj (obj) – instance of MCG class
bucket (obj) – MCG bucket to be used for workload
custom_options (dict) – Dictionary of lists containing tuples with additional configuration for fio in format: {‘section’: [(‘option’, ‘value’),…],…} e.g. {‘global’:[(‘name’,’bucketname’)],’create’:[(‘time_based’,’1’),(‘runtime’,’48h’)]} Those values can be added to the config or rewrite already existing values

Returns

Configmap definition

Return type

dict

ocs_ci.ocs.mcg_workload.get_job_dict(job_name)¶

Fio job dictionary with configuration set for MCG workload.

Parameters: job_name (str) – Name of the workload job
Returns: Specification for the workload job
Return type: dict

ocs_ci.ocs.mcg_workload.mcg_job_factory(request, bucket_factory, project_factory, mcg_obj, resource_path)¶

MCG IO workload factory. Calling this fixture creates a OpenShift Job.

Parameters

request (obj) – request fixture instance
bucket_factory (func) – factory function for bucket creation
project_factory (func) – factory function for project creation
mcg_obj (obj) – instance of MCG class
resource_path (str) – path to directory where should be created resources

Returns

MCG workload job factory function

Return type

func

ocs_ci.ocs.mcg_workload.wait_for_active_pods(job, desired_count, timeout=3)¶

Wait for job to load desired number of active pods in time specified in timeout.

Parameters

job (obj) – OCS job object
desired_count (str) – Number of desired active pods for provided job
timeout (int) – Number of seconds to wait for the job to get into state

Returns

If job has desired number of active pods

Return type

bool

ocs_ci.ocs.metrics module¶

OCS Metrics module.

Code in this module Supports monitoring test cases dealing with OCS metrics.

ocs_ci.ocs.metrics.get_missing_metrics(prometheus, metrics, current_platform=None)¶

Using given prometheus instance, check that all given metrics which are expected to be available on current platform are there.

Parameters

prometheus (ocs_ci.utility.prometheus.PrometheusAPI) – prometheus instance
metrics (list) – list or tuple with metrics to be checked
current_platform (str) – name of current platform (optional)

Returns

metrics which were not available but should be

Return type

list

ocs_ci.ocs.monitoring module¶

ocs_ci.ocs.monitoring.check_ceph_health_status_metrics_on_prometheus(mgr_pod)¶

Check ceph health status metric is collected on prometheus pod

Parameters: mgr_pod (str) – Name of the mgr pod
Returns: True on success, false otherwise
Return type: bool

ocs_ci.ocs.monitoring.check_ceph_metrics_available()¶

Check ceph metrics available

Returns: True on success, false otherwise
Return type: bool

ocs_ci.ocs.monitoring.check_pvcdata_collected_on_prometheus(pvc_name)¶

Checks whether initially pvc related data is collected on pod

Parameters: pvc_name (str) – Name of the pvc
Returns: True on success, raises UnexpectedBehaviour on failures

ocs_ci.ocs.monitoring.create_configmap_cluster_monitoring_pod(sc_name=None, telemeter_server_url=None)¶

Create a configmap named cluster-monitoring-config based on the arguments.

Parameters

sc_name (str) – Name of the storage class which will be used for persistent storage needs of OCP Prometheus and Alert Manager. If not defined, the related options won’t be present in the monitoring config map and the default (non persistent) storage will be used for OCP Prometheus and Alert Manager.
telemeter_server_url (str) – URL of Telemeter server where telemeter client (running in the cluster) will send it’s telemetry data. If not defined, related option won’t be present in the monitoring config map and the default (production) telemeter server will receive the metrics data.

ocs_ci.ocs.monitoring.get_list_pvc_objs_created_on_monitoring_pods()¶

Returns list of pvc objects created on monitoring pods

Returns: List of pvc objs
Return type: list

ocs_ci.ocs.monitoring.get_metrics_persistentvolumeclaims_info()¶

Returns the created pvc information on prometheus pod

Returns: The pvc metrics collected on prometheus pod
Return type: response.content (dict)

ocs_ci.ocs.monitoring.prometheus_health_check(name='monitoring', kind='ClusterOperator')¶

Return true if the prometheus cluster is healthy

Parameters

name (str) – Name of the resources
kind (str) – Kind of the resource

Returns

True on prometheus health is ok, false otherwise

Return type

bool

ocs_ci.ocs.monitoring.validate_pvc_are_mounted_on_monitoring_pods(pod_list)¶

Validate created pvc are mounted on monitoring pods

Parameters: pod_list (list) – List of the pods where pvc are mounted

ocs_ci.ocs.monitoring.validate_pvc_created_and_bound_on_monitoring_pods()¶

Validate pvc’s created and bound in state on monitoring pods

Raises: AssertionError – If no PVC are created or if any PVC are not in the Bound state

ocs_ci.ocs.node module¶

ocs_ci.ocs.node.add_new_node_and_label_it(machineset_name, num_nodes=1, mark_for_ocs_label=True)¶

Add a new node for ipi and label it

Parameters

machineset_name (str) – Name of the machine set
num_nodes (int) – number of nodes to add
mark_for_ocs_label (bool) – True if label the new node

eg: add_new_node_and_label_it(“new-tdesala-zlqzn-worker-us-east-2a”)

Returns: new spun node names
Return type: list

ocs_ci.ocs.node.add_new_node_and_label_upi(node_type, num_nodes, mark_for_ocs_label=True, node_conf=None)¶

Add a new node for aws/vmware upi platform and label it

Parameters

node_type (str) – Type of node, RHEL or RHCOS
num_nodes (int) – number of nodes to add
mark_for_ocs_label (bool) – True if label the new node
node_conf (dict) – The node configurations.

Returns

new spun node names

Return type

list

ocs_ci.ocs.node.add_new_node_to_lvd_and_lvs(old_node_name, new_node_name)¶

Replace the old node with the new node in localVolumeDiscovery and localVolumeSet, as described in the documents of node replacement with LSO

Parameters

old_node_name (str) – The old node name to remove from the local volume
new_node_name (str) – the new node name to add to the local volume

Returns

True in case if changes are applied. False otherwise

Return type

bool

ocs_ci.ocs.node.check_nodes_specs(min_memory, min_cpu)¶

Check that the cluster worker nodes meet the required minimum CPU and memory

Parameters

min_memory (int) – The required minimum memory in bytes
min_cpu (int) – The required minimum number of vCPUs

Returns

True if all nodes meet the required minimum specs, False otherwise

Return type

bool

ocs_ci.ocs.node.check_taint_on_ocs_nodes(taint='node.ocs.openshift.io/storage=true:NoSchedule')¶

Function to check for particular taint on nodes

Parameters: taint (str) – The taint to check on nodes
Returns: True if taint is present on node. False otherwise
Return type: bool

ocs_ci.ocs.node.delete_and_create_osd_node_aws_upi(osd_node_name)¶

Unschedule, drain and delete osd node, and creating a new osd node. At the end of the function there should be the same number of osd nodes as it was in the beginning, and also ceph health should be OK. This function is for AWS UPI.

Parameters: osd_node_name (str) – the name of the osd node
Returns: The new node name
Return type: str

ocs_ci.ocs.node.delete_and_create_osd_node_ipi(osd_node_name)¶

Unschedule, drain and delete osd node, and creating a new osd node. At the end of the function there should be the same number of osd nodes as it was in the beginning, and also ceph health should be OK.

This function is for any IPI platform.

Parameters: osd_node_name (str) – the name of the osd node
Returns: The new node name
Return type: str

ocs_ci.ocs.node.delete_and_create_osd_node_vsphere_upi(osd_node_name, use_existing_node=False)¶

Unschedule, drain and delete osd node, and creating a new osd node. At the end of the function there should be the same number of osd nodes as it was in the beginning, and also ceph health should be OK. This function is for vSphere UPI.

Parameters

osd_node_name (str) – the name of the osd node
use_existing_node (bool) – If False, create a new node and label it. If True, use an existing node to replace the deleted node and label it.

Returns

The new node name

Return type

str

ocs_ci.ocs.node.delete_and_create_osd_node_vsphere_upi_lso(osd_node_name, use_existing_node=False)¶

Unschedule, drain and delete osd node, and creating a new osd node. At the end of the function there should be the same number of osd nodes as it was in the beginning, and also ceph health should be OK. This function is for vSphere UPI.

Parameters

osd_node_name (str) – the name of the osd node
use_existing_node (bool) – If False, create a new node and label it. If True, use an existing node to replace the deleted node and label it.

Returns

The new node name

Return type

str

ocs_ci.ocs.node.drain_nodes(node_names)¶

Drain nodes

Parameters: node_names (list) – The names of the nodes
Raises: TimeoutExpired – in case drain command fails to complete in time

ocs_ci.ocs.node.get_all_nodes()¶

Gets the all nodes in cluster

Returns: List of node name
Return type: list

ocs_ci.ocs.node.get_app_pod_running_nodes(pod_obj)¶

Gets the app pod running node names

Parameters: pod_obj (list) – List of app pod objects
Returns: App pod running node names
Return type: list

ocs_ci.ocs.node.get_both_osd_and_app_pod_running_node(osd_running_nodes, app_pod_running_nodes)¶

Gets both osd and app pod running node names

Parameters

osd_running_nodes (list) – List of osd running node names
app_pod_running_nodes (list) – List of app pod running node names

Returns

Both OSD and app pod running node names

Return type

list

ocs_ci.ocs.node.get_compute_node_names(no_replace=False)¶

Gets the compute node names

Parameters: no_replace (bool) – If False ‘.’ will replaced with ‘-‘
Returns: List of compute node names
Return type: list

ocs_ci.ocs.node.get_master_nodes()¶

Fetches all master nodes.

Returns: List of names of master nodes
Return type: list

ocs_ci.ocs.node.get_node_az(node)¶

Get the node availability zone

Parameters: node (ocs_ci.ocs.resources.ocs.OCS) – The node object
Returns: The name of the node availability zone
Return type: str

ocs_ci.ocs.node.get_node_from_machine_name(machine_name)¶

Get node name from a given machine_name.

Parameters: machine_name (str) – Name of Machine
Returns: Name of Node (or None if not found)
Return type: str

ocs_ci.ocs.node.get_node_hostname_label(node_obj)¶

Get the hostname label of a node

Parameters: node_obj (ocs_ci.ocs.resources.ocs.OCS) – The node object
Returns: The node’s hostname label
Return type: str

ocs_ci.ocs.node.get_node_index_in_local_block(node_name)¶

Get the node index in the node values as it appears in the local block resource

Parameters: node_name (str) – The node name to search for his index
Returns: The node index in the nodeSelector values
Return type: int

ocs_ci.ocs.node.get_node_ips(node_type='worker')¶

Gets the node public IP

Parameters: node_type (str) – The node type (e.g. worker, master)
Returns: Node IP’s
Return type: list

ocs_ci.ocs.node.get_node_logs(node_name)¶

Get logs from a given node

pod_name (str): Name of the node

Returns: Output of ‘dmesg’ run on node
Return type: str

ocs_ci.ocs.node.get_node_name(node_obj)¶

Get oc node’s name

Parameters: node_obj (node_obj) – oc node object
Returns: node’s name
Return type: str

ocs_ci.ocs.node.get_node_objs(node_names=None)¶

Get node objects by node names

Parameters: node_names (list) – The node names to get their objects for. If None, will return all cluster nodes
Returns: Cluster node OCP objects
Return type: list

ocs_ci.ocs.node.get_node_pods(node_name, pods_to_search=None)¶

Get all the pods of a specified node

Parameters

node_name (str) – The node name to get the pods
pods_to_search (list) – list of pods to search for the node pods. If not specified, will search in all the pods.

Returns

list of all the pods of the specified node

Return type

list

ocs_ci.ocs.node.get_node_pods_to_scale_down(node_name)¶

Get the pods of a node to scale down as described in the documents of node replacement with LSO

Parameters: node_name (str) – The node name
Returns: The node’s pods to scale down
Return type: list

ocs_ci.ocs.node.get_node_resource_utilization_from_adm_top(nodename=None, node_type='worker', print_table=False)¶

Gets the node’s cpu and memory utilization in percentage using adm top command.

Parameters

nodename (str) – The node name
node_type (str) – The node type (e.g. master, worker)

Returns

Node name and its cpu and memory utilization in: percentage

Return type

dict

ocs_ci.ocs.node.get_node_resource_utilization_from_oc_describe(nodename=None, node_type='worker', print_table=False)¶

Gets the node’s cpu and memory utilization in percentage using oc describe node

Parameters

nodename (str) – The node name
node_type (str) – The node type (e.g. master, worker)

Returns

Node name and its cpu and memory utilization in: percentage

Return type

dict

ocs_ci.ocs.node.get_nodes(node_type='worker', num_of_nodes=None)¶

Get cluster’s nodes according to the node type (e.g. worker, master) and the number of requested nodes from that type

Parameters

node_type (str) – The node type (e.g. worker, master)
num_of_nodes (int) – The number of nodes to be returned

Returns

The nodes OCP instances

Return type

list

ocs_ci.ocs.node.get_ocs_nodes(num_of_nodes=None)¶

Gets the ocs nodes

Parameters: num_of_nodes (int) – The number of ocs nodes to return. If not specified, it returns all the ocs nodes.
Returns: List of ocs nodes
Return type: list

ocs_ci.ocs.node.get_osd_running_nodes()¶

Gets the osd running node names

Returns: OSD node names
Return type: list

ocs_ci.ocs.node.get_osds_per_node()¶

Gets the osd running pod names per node name

Returns: {“Node name”:[“osd running pod name running on the node”,..,]}
Return type: dict

ocs_ci.ocs.node.get_provider()¶

Return the OCP Provider (Platform)

Returns: The Provider that the OCP is running on
Return type: str

ocs_ci.ocs.node.get_running_pod_count_from_node(nodename=None, node_type='worker')¶

Gets the node running pod count using oc describe node

Parameters

nodename (str) – The node name
node_type (str) – The node type (e.g. master, worker)

Returns

Node name and its pod_count

Return type

dict

ocs_ci.ocs.node.get_typed_worker_nodes(os_id='rhcos')¶

Get worker nodes with specific OS

Parameters: os_id (str) – OS type like rhcos, RHEL etc…
Returns: list of worker nodes instances having specified os
Return type: list

ocs_ci.ocs.node.get_worker_nodes()¶

Fetches all worker nodes.

Returns: List of names of worker nodes
Return type: list

ocs_ci.ocs.node.get_worker_nodes_not_in_ocs()¶

Get the worker nodes that are not ocs labeled.

Returns: list of worker node objects that are not ocs labeled
Return type: list

ocs_ci.ocs.node.is_node_labeled(node_name, label="cluster.ocs.openshift.io/openshift-storage=''")¶

Check if the node is labeled with a specified label.

Parameters

node_name (str) – The node name to check if it has the specific label
label (str) – The name of the label. Default value is the OCS label.

Returns

True if the node is labeled with the specified label. False otherwise

Return type

bool

ocs_ci.ocs.node.label_nodes(nodes, label="cluster.ocs.openshift.io/openshift-storage=''")¶

Label nodes

Parameters

nodes (list) – list of node objects need to label
label (str) – New label to be assigned for these nodes. Default value is the OCS label

ocs_ci.ocs.node.node_network_failure(node_names, wait=True)¶

Induce node network failure Bring node network interface down, making the node unresponsive

Parameters

node_names (list) – The names of the nodes
wait (bool) – True in case wait for status is needed, False otherwise

Returns

True if node network fail is successful

Return type

bool

ocs_ci.ocs.node.node_replacement_verification_steps_ceph_side(old_node_name, new_node_name, new_osd_node_name)¶

Check the verification steps from the Ceph side, after the process of node replacement as described in the docs

Parameters

old_node_name (str) – The name of the old node that has been deleted
new_node_name (str) – The name of the new node that has been created
new_osd_node_name (str) – The name of the new node that has been added to osd nodes

Returns

True if all the verification steps passed. False otherwise

Return type

bool

ocs_ci.ocs.node.node_replacement_verification_steps_user_side(old_node_name, new_node_name, new_osd_node_name, old_osd_id)¶

Check the verification steps that the user should perform after the process of node replacement as described in the docs

Parameters

old_node_name (str) – The name of the old node that has been deleted
new_node_name (str) – The name of the new node that has been created
new_osd_node_name (str) – The name of the new node that has been added to osd nodes
old_osd_id (str) – The old osd id

Returns

True if all the verification steps passed. False otherwise

Return type

bool

ocs_ci.ocs.node.print_table_node_resource_utilization(utilization_dict, field_names)¶

Print table of node utilization

Parameters

utilization_dict (dict) – CPU and Memory utilization per Node
field_names (list) – The field names of the table

ocs_ci.ocs.node.remove_nodes(nodes)¶

Remove the nodes from cluster

Parameters: nodes (list) – list of node instances to remove from cluster

ocs_ci.ocs.node.scale_down_deployments(node_name)¶

Scale down the deployments of a node as described in the documents of node replacement with LSO

Parameters: node_name (str) – The node name

ocs_ci.ocs.node.schedule_nodes(node_names)¶

Change nodes to be scheduled

Parameters: node_names (list) – The names of the nodes

ocs_ci.ocs.node.taint_nodes(nodes, taint_label='node.ocs.openshift.io/storage')¶

Taint nodes

Parameters

nodes (list) – list of node names need to taint
taint_label (str) – New taint label to be assigned for these nodes. Default value is the OCS taint

ocs_ci.ocs.node.taint_ocs_nodes(nodes_to_taint=None)¶

Function to taint nodes with “node.ocs.openshift.io/storage=true:NoSchedule”

Parameters: nodes_to_taint (list) – Nodes to taint

ocs_ci.ocs.node.unschedule_nodes(node_names)¶

Change nodes to be unscheduled

Parameters: node_names (list) – The names of the nodes

ocs_ci.ocs.node.untaint_ocs_nodes(taint='node.ocs.openshift.io/storage=true:NoSchedule', nodes_to_untaint=None)¶

Function to remove taints from nodes

Parameters

taint (str) – taint to use
nodes_to_taint (list) – list of nodes to untaint

Returns

True if untainted, false otherwise

Return type

bool

ocs_ci.ocs.node.wait_for_new_osd_node(old_osd_node_names, timeout=180)¶

Wait for the new osd node to appear.

Parameters

old_osd_node_names (list) – List of the old osd node names
timeout (int) – time to wait for the new osd node to appear

Returns

The new osd node name if the new osd node appear in the specific timeout.: Else it returns None

Return type

str

ocs_ci.ocs.node.wait_for_nodes_status(node_names=None, status='Ready', timeout=180)¶

Wait until all nodes are in the given status

Parameters

node_names (list) – The node names to wait for to reached the desired state If None, will wait for all cluster nodes
status (str) – The node status to wait for (e.g. ‘Ready’, ‘NotReady’, ‘SchedulingDisabled’)
timeout (int) – The number in seconds to wait for the nodes to reach the status

Raises

ResourceWrongStatusException – In case one or more nodes haven’t reached the desired state

ocs_ci.ocs.ocp module¶

General OCP object

class ocs_ci.ocs.ocp.OCP(api_version='v1', kind='Service', namespace=None, resource_name='', selector=None)¶

Bases: object

A basic OCP object to run basic ‘oc’ commands

add_label(resource_name, label)¶

Adds a new label for this pod

Parameters

resource_name (str) – Name of the resource you want to label
label (str) – New label to be assigned for this pod E.g: “label=app=’rook-ceph-mds’”

property api_version¶

apply(yaml_file)¶

Applies configuration changes to a resource

Parameters: yaml_file (str) – Path to a yaml file to use in ‘oc apply -f file.yaml
Returns: Dictionary represents a returned yaml file
Return type: dict

check_function_supported(support_var)¶

Check if the resource supports the functionality based on the support_var.

Parameters: support_var (bool) – True if functionality is supported, False otherwise.
Raises: NotSupportedFunctionError – If support_var == False

check_name_is_specified(resource_name='')¶

Check if the name of the resource is specified in class level and if not raise the exception.

Raises: ResourceNameNotSpecifiedException – in case the name is not specified.

check_phase(phase)¶

Check phase of resource

Parameters

phase (str) – Phase of resource object

Returns

True if phase of object is the same as passed one, False: otherwise.

Return type

bool

Raises

NotSupportedFunctionError – If resource doesn’t have phase!
ResourceNameNotSpecifiedException – in case the name is not specified.

create(yaml_file=None, resource_name='', out_yaml_format=True)¶

Creates a new resource

Parameters

yaml_file (str) – Path to a yaml file to use in ‘oc create -f file.yaml
resource_name (str) – Name of the resource you want to create
out_yaml_format (bool) – Determines if the output should be formatted to a yaml like string

Returns

Dictionary represents a returned yaml file

Return type

dict

property data¶

delete(yaml_file=None, resource_name='', wait=True, force=False)¶

Deletes a resource

Parameters

yaml_file (str) – Path to a yaml file to use in ‘oc delete -f file.yaml
resource_name (str) – Name of the resource you want to delete
wait (bool) – Determines if the delete command should wait to completion
force (bool) – True for force deletion with –grace-period=0, False otherwise

Returns

Dictionary represents a returned yaml file

Return type

dict

Raises

CommandFailed – In case yaml_file and resource_name wasn’t provided

delete_project(project_name)¶

Delete a project. A project created by the new_project function does not have a corresponding yaml file so normal resource deletion calls do not work

Parameters: project_name (str) – Name of the project to be deleted
Returns: True in case project deletion succeeded.
Return type: bool
Raises: CommandFailed – When the project deletion does not succeed.

describe(resource_name='', selector=None, all_namespaces=False)¶

Get command - ‘oc describe <resource>’

Parameters

resource_name (str) – The resource name to fetch
selector (str) – The label selector to look for
all_namespaces (bool) – Equal to oc describe <resource> -A

Example

describe(‘my-pv1’)

Returns: Dictionary represents a returned yaml file
Return type: dict

exec_oc_cmd(command, out_yaml_format=True, secrets=None, timeout=600, ignore_error=False, **kwargs)¶

Executing ‘oc’ command

Parameters

command (str) – The command to execute (e.g. create -f file.yaml) without the initial ‘oc’ at the beginning
out_yaml_format (bool) – whether to return yaml loaded python object or raw output
secrets (list) – A list of secrets to be masked with asterisks This kwarg is popped in order to not interfere with subprocess.run(**kwargs)
timeout (int) – timeout for the oc_cmd, defaults to 600 seconds
ignore_error (bool) – True if ignore non zero return code and do not raise the exception.

Returns

Dictionary represents a returned yaml file. str: If out_yaml_format is False.

Return type

dict

exec_oc_debug_cmd(node, cmd_list, timeout=300)¶

Function to execute “oc debug” command on OCP node

Parameters

node (str) – Node name where the command to be executed
cmd_list (list) – List of commands eg: [‘cmd1’, ‘cmd2’]
timeout (int) – timeout for the exec_oc_cmd, defaults to 600 seconds

Returns

Returns output of the executed command/commands

Return type

out (str)

Raises

CommandFailed – When failure in command execution

get(resource_name='', out_yaml_format=True, selector=None, all_namespaces=False, retry=0, wait=3, dont_raise=False)¶

Get command - ‘oc get <resource>’

Parameters

resource_name (str) – The resource name to fetch
out_yaml_format (bool) – Adding ‘-o yaml’ to oc command
selector (str) – The label selector to look for.
all_namespaces (bool) – Equal to oc get <resource> -A
retry (int) – Number of attempts to retry to get resource
wait (int) – Number of seconds to wait between attempts for retry
dont_raise (bool) – If True will raise when get is not found

Example

get(‘my-pv1’)

Returns: Dictionary represents a returned yaml file None: Incase dont_raise is True and get is not found
Return type: dict

get_logs(name, container_name=None, all_containers=False, secrets=None, timeout=None, ignore_error=False)¶

Execute oc logs command to fetch logs for a given k8s resource.

Since the log is stored as a string in memory, this will be problematic when the log is large.

Parameters

name (str) – name of the resource to fetch logs from
container_name (str) – name of the container (optional)
all_containers (bool) – fetch logs from all containers of the resource
secrets (list) – A list of secrets to be masked with asterisks
timeout (int) – timeout for the oc_cmd
ignore_error (bool) – True if ignore non zero return code and do not raise the exception.

Returns

container logs

Return type

str

get_resource(resource_name, column, retry=0, wait=3, selector=None)¶

Get a column value for a resource based on: ‘oc get <resource_kind> <resource_name>’ command

Parameters

resource_name (str) – The name of the resource to get its column value
column (str) – The name of the column to retrive
retry (int) – Number of attempts to retry to get resource
wait (int) – Number of seconds to wait beteween attempts for retry
selector (str) – The resource selector to search with.

Returns

The output returned by ‘oc get’ command not in the ‘yaml’: format

Return type

str

get_resource_status(resource_name)¶

Get the resource STATUS column based on: ‘oc get <resource_kind> <resource_name>’ command

Parameters

resource_name (str) – The name of the resource to get its STATUS

Returns

The status returned by ‘oc get’ command not in the ‘yaml’: format

Return type

str

get_user_token()¶

Get user access token

Returns: access token
Return type: str

is_exist(resource_name='', selector=None)¶

Check if at least one of the resource exists.

Parameters

resource_name (str) – Name of the resource.
selector (str) – Selector of the resource.

Raises

ResourceNameNotSpecifiedException – In case the name is not specified.

Returns

True if the resource exists False otherwise.

Return type

bool

property kind¶

login(user, password)¶

Logs user in

Parameters

user (str) – Name of user to be logged in
password (str) – Password of user to be logged in

Returns

output of login command

Return type

str

login_as_sa()¶

Logs in as system:admin

Returns: output of login command
Return type: str

property namespace¶

new_project(project_name)¶

Creates a new project

Parameters: project_name (str) – Name of the project to be created
Returns: True in case project creation succeeded, False otherwise
Return type: bool

patch(resource_name='', params=None, format_type='')¶

Applies changes to resources

Parameters

resource_name (str) – Name of the resource
params (str) – Changes to be added to the resource
type (str) – Type of the operation

Returns

True in case if changes are applied. False otherwise

Return type

bool

reload_data()¶: Reloading data of OCP object

property resource_name¶

wait_for_delete(resource_name='', timeout=60, sleep=3)¶

Wait for a resource to be deleted

Parameters

resource_name (str) – The name of the resource to wait for (e.g.my-pv1)
timeout (int) – Time in seconds to wait
sleep (int) – Sampling time in seconds

Raises

CommandFailed – If failed to verify the resource deletion
TimeoutError – If resource is not deleted within specified timeout

Returns

True in case resource deletion is successful

Return type

bool

wait_for_phase(phase, timeout=300, sleep=5)¶

Wait till phase of resource is the same as required one passed in the phase parameter.

Parameters

phase (str) – Desired phase of resource object
timeout (int) – Timeout in seconds to wait for desired phase
sleep (int) – Time in seconds to sleep between attempts

Raises

ResourceInUnexpectedState – In case the resource is not in expected phase.
NotSupportedFunctionError – If resource doesn’t have phase!
ResourceNameNotSpecifiedException – in case the name is not specified.

wait_for_resource(condition, resource_name='', column='STATUS', selector=None, resource_count=0, timeout=60, sleep=3, dont_allow_other_resources=False, error_condition=None)¶

Wait for a resource to reach to a desired condition

Parameters

condition (str) – The desired state the resource that is sampled from ‘oc get <kind> <resource_name>’ command
resource_name (str) – The name of the resource to wait for (e.g.my-pv1)
column (str) – The name of the column to compare with
selector (str) – The resource selector to search with. Example: ‘app=rook-ceph-mds’
resource_count (int) – How many resources expected to be
timeout (int) – Time in seconds to wait
sleep (int) – Sampling time in seconds
dont_allow_other_resources (bool) – If True it will not allow other resources in different state. For example you are waiting for 2 resources and there are currently 3 (2 in running state, 1 in ContainerCreating) the function will continue to next iteration to wait for only 2 resources in running state and no other exists.
error_condition (str) – State of the resource that is sampled from ‘oc get <kind> <resource_name>’ command, which makes this method to fail immediately without waiting for a timeout. This is optional and makes sense only when there is a well defined unrecoverable state of the resource(s) which is not expected to be part of a workflow under test, and at the same time, the timeout itself is large.

Returns

True in case all resources reached desired condition,: False otherwise

Return type

bool

ocs_ci.ocs.ocp.confirm_cluster_operator_version(target_version, cluster_operator)¶

Check if cluster operator upgrade process is completed:

Parameters

cluster_operator – (str): ClusterOperator name
target_version (str) – expected OCP client

Returns

True if success, False if failed

Return type

bool

ocs_ci.ocs.ocp.get_all_cluster_operators()¶

Get all ClusterOperators names in OCP

Returns: cluster-operator names
Return type: list

ocs_ci.ocs.ocp.get_build()¶

Return the OCP Build Version

Returns: The build version of the OCP
Return type: str

ocs_ci.ocs.ocp.get_cluster_operator_version(cluster_operator_name)¶

Get image version of selected cluster operator

Parameters: cluster_operator_name (str) – ClusterOperator name
Returns: cluster operator version: ClusterOperator image version
Return type: str

ocs_ci.ocs.ocp.get_clustername()¶

Return the name (DNS short name) of the cluster

Returns: the short DNS name of the cluster
Return type: str

ocs_ci.ocs.ocp.get_current_oc_version()¶

Gets Current OCP client version

Returns: current COP client version
Return type: str

ocs_ci.ocs.ocp.get_images(data, images=None)¶

Get the images from the ocp object like pod, CSV and so on.

Parameters

data (dict) – Yaml data from the object.
images (dict) – Dict where to put the images (doesn’t have to be set!).

Returns

Images dict like: {‘image_name’: ‘image.url.to:tag’, …}

Return type

dict

ocs_ci.ocs.ocp.get_ocp_channel()¶

Return the OCP Channel

Returns: The channel of the OCP
Return type: str

ocs_ci.ocs.ocp.get_ocp_upgrade_channel()¶

Gets OCP upgrade channel

Returns: OCP upgrade channel name
Return type: str

ocs_ci.ocs.ocp.get_ocs_parsed_version()¶

Returns ocs version as float

Returns: ocs version number as major.minor (for example: 4.5)
Return type: float

ocs_ci.ocs.ocp.get_ocs_version()¶

Return the OCS Version

Returns: The version of the OCS
Return type: str

ocs_ci.ocs.ocp.patch_ocp_upgrade_channel(channel_variable='candidate-4.7')¶

Using ‘oc patch clusterversion’ if new OCP upgrade channel is different than current one

Parameters: channel_variable (str) – New OCP upgrade subscription channel

ocs_ci.ocs.ocp.rsync(src, dst, node, dst_node=True, extra_params='')¶

This function will rsync source folder to destination path. You can rsync local folder to the node or vice versa depends on dst_node parameter. By default the rsync is from local to the node.

Parameters

src (str) – Source path of folder to rsync.
dst (str) – Destination path where to rsync.
node (str) – Node to/from copy.
dst_node (bool) – True if the destination (dst) is the node, False when dst is the local folder.
extra_params (str) – “See: oc rsync –help for the extra params”

ocs_ci.ocs.ocp.switch_to_default_rook_cluster_project()¶

Switch to default project

Returns: True on success, False otherwise
Return type: bool

ocs_ci.ocs.ocp.switch_to_project(project_name)¶

Switch to another project

Parameters: project_name (str) – Name of the project to be switched to
Returns: True on success, False otherwise
Return type: bool

ocs_ci.ocs.ocp.upgrade_ocp(image_path, image)¶

upgrade OCP version

Parameters

image (str) – image to be installed
image_path (str) – path to image

ocs_ci.ocs.ocp.validate_cluster_version_status()¶

Verify OCP upgrade is completed, by checking ‘oc get clusterversion’ status

Returns

False in case that one of condition flags is invalid:: Progressing (should be False), Failing(should be False) or Available (should be True)

Return type

bool

ocs_ci.ocs.ocp.verify_cluster_operator_status(cluster_operator)¶

Checks if cluster operator status is degraded or progressing, as sign that upgrade not yet completed

Parameters: cluster_operator (str) – OCP cluster operator name
Returns: True if cluster operator status is valid, False if cluster operator status is “degraded” or “progressing”
Return type: bool

ocs_ci.ocs.ocp.verify_images_upgraded(old_images, object_data)¶

Verify that all images in ocp object are upgraded.

Parameters

old_images (set) – Set with old images.
object_data (dict) – OCP object yaml data.

Raises

NonUpgradedImagesFoundError – In case the images weren’t upgraded.

ocs_ci.ocs.ocp.verify_ocp_upgrade_channel(channel_variable='candidate-4.7')¶

When upgrade OCP version, verify that subscription channel is same as current one

Parameters

channel_variable (str) – New OCP upgrade subscription channel

Returns

True when OCP subscription channel is correct,: and no patch needed

Return type

bool

ocs_ci.ocs.ocp.wait_for_cluster_connectivity(tries=200, delay=3)¶

Wait for the cluster to be reachable

Parameters

tries (int) – The number of retries
delay (int) – The delay in seconds between retries

Returns

True if cluster is reachable, False otherwise

Return type

bool

Raises

CommandFailed – In case the cluster is unreachable

ocs_ci.ocs.ocs_upgrade module¶

class ocs_ci.ocs.ocs_upgrade.OCSUpgrade(namespace, version_before_upgrade, ocs_registry_image, upgrade_in_current_source)¶

Bases: object

OCS Upgrade helper class

check_if_upgrade_completed(channel, csv_name_pre_upgrade)¶

Checks if OCS operator finishes it’s upgrade

Parameters

channel – (str): OCS subscription channel
csv_name_pre_upgrade – (str): OCS operator name

Returns

True if upgrade completed, False otherwise

Return type

bool

get_csv_name_pre_upgrade()¶

Getting OCS operator name as displayed in CSV

Returns: OCS operator name, as displayed in CSV
Return type: str

get_images_post_upgrade(channel, pre_upgrade_images, upgrade_version)¶

Checks if all images of OCS cluster upgraded,: and return list of all images if upgrade success

Parameters

channel – (str): OCS subscription channel
pre_upgrade_images – (dict): Contains all OCS cluster images
upgrade_version – (str): version to be upgraded

Returns

Contains full path of OCS cluster old images

Return type

set

get_parsed_versions()¶

Getting parsed version names, current version, and upgrade version

Returns

contains 2 strings with parsed names of current version: and upgrade version

Return type

tuple

get_pre_upgrade_image(csv_name_pre_upgrade)¶

Getting all OCS cluster images, before upgrade

Parameters

csv_name_pre_upgrade – (str): OCS operator name

Returns

Contains all OCS cluster images: dict keys: Image name dict values: Image full path

Return type

dict

get_upgrade_version()¶

Getting the required upgrade version

Returns: version to be upgraded
Return type: str

load_version_config_file(upgrade_version)¶

Loads config file to the ocs-ci config with upgrade version

Parameters: upgrade_version – (str): version to be upgraded

property ocs_registry_image¶

set_upgrade_channel()¶

Wait for the new package manifest for upgrade.

Returns: OCS subscription channel
Return type: str

set_upgrade_images()¶: Set images for upgrade

update_subscription(channel)¶

Updating OCS operator subscription

Parameters: channel – (str): OCS subscription channel

property version_before_upgrade¶

ocs_ci.ocs.ocs_upgrade.get_upgrade_image_info(old_csv_images, new_csv_images)¶

Log the info about the images which are going to be upgraded.

Parameters

old_csv_images (dict) – dictionary with the old images from pre upgrade CSV
new_csv_images (dict) – dictionary with the post upgrade images

Returns

with three sets which contain those types of images: old_images_for_upgrade - images which should be upgraded new_images_to_upgrade - new images to which upgrade unchanged_images - unchanged images without upgrade

Return type

tuple

ocs_ci.ocs.ocs_upgrade.run_ocs_upgrade(operation=None, *operation_args, **operation_kwargs)¶

Run upgrade procedure of OCS cluster

Parameters

operation – (function): Function to run
operation_args – (iterable): Function’s arguments
operation_kwargs – (map): Function’s keyword arguments

ocs_ci.ocs.ocs_upgrade.verify_image_versions(old_images, upgrade_version, version_before_upgrade)¶

Verify if all the images of OCS objects got upgraded

Parameters

old_images (set) – set with old images
upgrade_version (packaging.version.Version) – version of OCS
version_before_upgrade (float) – version of OCS before upgrade

ocs_ci.ocs.openshift_ops module¶

class ocs_ci.ocs.openshift_ops.OCP¶

Bases: object

Class which contains various utility functions for interacting with OpenShift

static call_api(method, **kw)¶

This function makes generic REST calls

Parameters

method (str) – one of the GET, CREATE, PATCH, POST, DELETE
**kw – Based on context of the call kw will be populated by caller

Returns

ResourceInstance object

create_project(project)¶

Creates new project

Parameters: project (str) – project name
Returns: True if successful otherwise False
Return type: bool

get_labels(pod_name, pod_namespace)¶

Get labels from specific pod

Parameters

pod_name (str) – Name of pod in oc cluster
pod_namespace (str) – pod namespace in which the pod lives

Raises

NotFoundError – If resource not found

Returns

All the openshift labels on a given pod

Return type

dict

get_pods(**kw)¶

Get pods in specific namespace or across oc cluster.

Parameters

**kw – ex: namespace=rook-ceph, label_selector=’x==y’

Returns

of pods names, if no namespace provided then this function: returns all pods across openshift cluster.

Return type

list

get_projects()¶

Gets all the projects in the cluster

Returns: List of projects
Return type: list

get_services()¶

Gets all the services in the cluster

Returns

defaultdict of services, key represents the namespace: and value represents the services

Return type

dict

get_services_in_namespace(namespace)¶

Gets the services in a namespace

Returns: list of services in a namespace
Return type: list

static set_kubeconfig(kubeconfig_path)¶

Export environment variable KUBECONFIG for future calls of OC commands or other API calls

Parameters: kubeconfig_path (str) – path to kubeconfig file to be exported
Returns: True if successfully connected to cluster, False otherwise
Return type: boolean

ocs_ci.ocs.openstack module¶

class ocs_ci.ocs.openstack.CephVMNode(**kw)¶

Bases: object

attach_floating_ip(timeout=120)¶

create_node(**kw)¶

destroy_node()¶: Relies on the fact that names should be unique. Along the chain we prevent non-unique names to be used/added. TODO: raise an exception if more than one node is matched to the name, that can be propagated back to the client.

destroy_volume(name)¶

get_driver(**kw)¶

get_private_ip()¶: Workaround. self.node.private_ips returns empty list.

get_volume(name)¶: Return libcloud.compute.base.StorageVolume

exception ocs_ci.ocs.openstack.GetIPError¶: Bases: Exception

exception ocs_ci.ocs.openstack.InvalidHostName¶: Bases: Exception

exception ocs_ci.ocs.openstack.NodeErrorState¶: Bases: Exception

ocs_ci.ocs.parallel module¶

class ocs_ci.ocs.parallel.ExceptionHolder(exc_info)¶: Bases: object

ocs_ci.ocs.parallel.capture_traceback(func, *args, **kwargs)¶: Utility function to capture tracebacks of any exception func raises.

class ocs_ci.ocs.parallel.parallel¶

Bases: object

This class is a context manager for running functions in parallel.

You add functions to be run with the spawn method:

with parallel() as p:
    for foo in bar:
        p.spawn(quux, foo, baz=True)

You can iterate over the results (which are in arbitrary order):

with parallel() as p:
    for foo in bar:
        p.spawn(quux, foo, baz=True)
    for result in p:
        print result

If one of the spawned functions throws an exception, it will be thrown when iterating over the results, or when the with block ends.

At the end of the with block, the main thread waits until all spawned functions have completed, or, if one exited with an exception, kills the rest and raises the exception.

spawn(func, *args, **kwargs)¶

ocs_ci.ocs.parallel.resurrect_traceback(exc)¶

ocs_ci.ocs.perfresult module¶

Basic Module to manage performance results

class ocs_ci.ocs.perfresult.PerfResult(uuid, crd)¶

Bases: object

Basic Performance results object for Q-PAS team

add_key(key, value)¶

Adding (key and value) to this object results dictionary as a new dictionary.

Parameters

key (str) – String which will be the key for the value
value – value to add, can be any kind of data type

es_connect()¶: Create Elastic-Search server connection

es_read()¶

Reading all test results from the elastic-search server

Returns: list of results
Return type: list

Assert:: if no data found in the server

es_write()¶

Writing the results to the elastic-search server

Raise:: RequestError: in case of error writing data to the server

results_link()¶

Create a link to the results of the test in the elasticsearch serer

Returns: http link to the test results in the elastic-search server
Return type: str

ocs_ci.ocs.perftests module¶

class ocs_ci.ocs.perftests.PASTest¶

Bases: ocs_ci.framework.testlib.BaseTest

Base class for QPAS team - Performance and Scale tests

This class contain functions which used by performance and scale test, and also can be used by E2E test which used the benchmark-operator (ripsaw)

copy_es_data(elasticsearch)¶

Copy data from Internal ES (if exists) to the main ES

Parameters: elasticsearch (obj) – elasticsearch object (if exits)

deploy_and_wait_for_wl_to_start(timeout=300, sleep=20)¶

Deploy the workload and wait until it start working

Parameters

timeout (int) – time in second to wait until the benchmark start
sleep (int) – Sleep interval seconds

es_info_backup(elasticsearch)¶

Saving the Original elastic-search IP and PORT - if defined in yaml

Parameters: elasticsearch (obj) – elasticsearch object

get_env_info()¶: Getting the environment information and update the workload RC if necessary.

pytestmark = [Mark(name='usefixtures', args=('environment_checker',), kwargs={}), Mark(name='performance', args=(), kwargs={}), Mark(name='scale', args=(), kwargs={})]¶

ripsaw_deploy(ripsaw)¶

Deploy the benchmark operator (formally ripsaw) CRD

Parameters: ripsaw (obj) – benchmark operator object

set_storageclass(interface)¶

Setting the benchmark CRD storageclass

Parameters: interface (str) – The interface which will used in the test

setup()¶: Setting up the environment for each performance and scale test

wait_for_wl_to_finish(timeout=18000, sleep=300)¶

Waiting until the workload is finished and get the test log

Parameters

timeout (int) – time in second to wait until the benchmark start
sleep (int) – Sleep interval seconds

ocs_ci.ocs.pgsql module¶

Postgresql workload class

class ocs_ci.ocs.pgsql.Postgresql(**kwargs)¶

Bases: ocs_ci.ocs.ripsaw.RipSaw

Postgresql workload operation

attach_pgsql_pod_to_claim_pvc(pvc_objs, postgres_name, run_benchmark=True, pgbench_name=None)¶

Attaches pgsql pod to created claim PVC

Parameters

pvc_objs (list) – List of PVC objs which needs to attached to pod
postgres_name (str) – Name of the postgres pod
run_benchmark (bool) – On true, runs pgbench benchmark on postgres pod
pgbench_name (str) – Name of pgbench benchmark

Returns

List of pod objs created

Return type

pgsql_obj_list (list)

cleanup()¶: Clean up

create_pgbench_benchmark(replicas, pgbench_name=None, postgres_name=None, clients=None, threads=None, transactions=None, scaling_factor=None, timeout=None, wait=True)¶

Create pgbench benchmark pods

Parameters

replicas (int) – Number of pgbench pods to be deployed
pgbench_name (str) – Name of pgbench bechmark
postgres_name (str) – Name of postgres pod
clients (int) – Number of clients
threads (int) – Number of threads
transactions (int) – Number of transactions
scaling_factor (int) – scaling factor
timeout (int) – Time in seconds to wait
wait (bool) – On true waits till pgbench reaches Completed state

Returns

pgbench pod objects list

Return type

List

delete_pgbench_pods(pg_obj_list)¶

Delete all pgbench pods on cluster

Returns: True if deleted, False otherwise
Return type: bool

export_pgoutput_to_googlesheet(pg_output, sheet_name, sheet_index)¶

Collect pgbench output to google spreadsheet

Parameters

pg_output (list) – pgbench outputs in list
sheet_name (str) – Name of the sheet
sheet_index (int) – Index of sheet

get_pgbech_pod_status_table(pgbench_pods)¶

Get pgbench pod data and print results on a table

Parameters: pods (pgbench) – List of pgbench pods

get_pgbench_pods()¶

Get all pgbench pods

Returns: pgbench pod objects list
Return type: List

get_pgbench_status(pgbench_pod_name)¶

Get pgbench status

Parameters: pgbench_pod_name (str) – Name of the pgbench pod
Returns: state of pgbench pod (running/completed)
Return type: str

get_pgsql_nodes()¶

Get nodes that contain a pgsql app pod

Returns: Cluster node OCP objects
Return type: list

get_postgres_pods()¶: Get all postgres pods :returns: postgres pod objects list :rtype: List

get_postgres_pvc()¶

Get all postgres pvc

Returns: postgres pvc objects list
Return type: List

get_postgres_used_file_space(pod_obj_list)¶

Get the used file space on a mount point

Parameters: pod_obj_list (POD) – List of pod objects
Returns: List of pod object
Return type: list

is_pgbench_running()¶

Check if pgbench is running

Returns: True if pgbench is running; False otherwise
Return type: bool

respin_pgsql_app_pod()¶

Respin the pgsql app pod

Returns: pod status

setup_postgresql(replicas, sc_name=None)¶

Deploy postgres sql server

Parameters: replicas (int) – Number of postgresql pods to be deployed
Raises: CommandFailed – If PostgreSQL server setup fails

validate_pgbench_run(pgbench_pods, print_table=True)¶

Validate pgbench run

Parameters: pods (pgbench) – List of pgbench pods
Returns: pgbench outputs in list
Return type: pg_output (list)

wait_for_pgbench_status(status, timeout=None)¶

Wait for pgbench benchmark pods status to reach running/completed

Parameters

status (str) – status to reach Running or Completed
timeout (int) – Time in seconds to wait

wait_for_postgres_status(status='Running', timeout=300)¶

Wait for postgres pods status to reach running/completed

Parameters

status (str) – status to reach Running or Completed
timeout (int) – Time in seconds to wait

ocs_ci.ocs.pillowfight module¶

Pillowfight Class to run various workloads and scale tests

class ocs_ci.ocs.pillowfight.PillowFight(**kwargs)¶

Bases: object

Workload operation using PillowFight This class was modelled after the RipSaw class in this directory.

MAX_ACCEPTABLE_RESPONSE_TIME = 2000¶

MIN_ACCEPTABLE_OPS_PER_SEC = 2000¶

WAIT_FOR_TIME = 1800¶

analyze_all()¶: Analyze the data extracted into self.logs files

cleanup()¶: Remove pillowfight pods and temp files

export_pfoutput_to_googlesheet(sheet_name, sheet_index)¶

Collect pillowfight output to google spreadsheet

Parameters

sheet_name (str) – Name of the sheet
sheet_index (int) – Index of sheet

parse_pillowfight_log(data_from_log)¶

Run oc logs on the pillowfight pod passed in. Cleanup the output from oc logs to handle peculiarities in the couchbase log results, and generate a summary of the results.

The dictionary returned has two values; ‘opspersec’ and ‘resptimes’. opspersec is a list of ops per second numbers reported.’ resptimes is a dictionary index by the max response time of a range. Each entry in resptimes contains a minimum response time for that range, and a count of how many messages fall within that range.

Parameters: data_from_log (str) – log data
Returns: ops per sec and response time information
Return type: dict

run_pillowfights(replicas=1, num_items=None, num_threads=None)¶

loop through all the yaml files extracted from the pillowfight repo and run them. Run oc logs on the results and save the logs in self.logs directory

Parameters

replicas (int) – Number of pod replicas
num_items (int) – Number of items to be loaded to the cluster
num_threads (int) – Number of threads

sanity_check(stats)¶: Make sure the worst cases for ops per second and response times are within an acceptable range.

ocs_ci.ocs.platform_nodes module¶

class ocs_ci.ocs.platform_nodes.AWSNodes¶

Bases: ocs_ci.ocs.platform_nodes.NodesBase

AWS EC2 instances class

all_nodes_found(node_names)¶

Relying on oc get nodes -o wide to confirm that node is added to cluster

Parameters: node_names (list) – of node names as string
Returns: ‘True’ if all the node names appeared in ‘get nodes’ else ‘False’
Return type: bool

approve_all_nodes_csr(node_list)¶

Make sure that all the newly added nodes are in approved csr state

Parameters: node_list (list) – of AWSUPINode/AWSIPINode objects
Raises: PendingCSRException – If any pending csrs exists

attach_nodes_to_cluster(node_list)¶

Attach nodes in the list to the cluster

Parameters: node_list (list) – of AWSUPINode/AWSIPINode objects

attach_nodes_to_upi_cluster(node_list)¶

Attach node to upi cluster Note: For RHCOS nodes, create function itself would have attached the nodes to cluster so nothing to do here

Parameters: node_list (list) – of AWSUPINode objects

attach_rhel_nodes_to_upi_cluster(node_list)¶

Attach RHEL nodes to upi cluster

Parameters: node_list (list) – of AWSUPINode objects with RHEL os

attach_volume(volume, node)¶

Attach a data volume to an instance

Parameters

volume (Volume) – The volume to delete
node (OCS) – The EC2 instance to attach the volume to

build_ansible_inventory(hosts)¶

Build the ansible hosts file from jinja template

Parameters: hosts (list) – list of private host names
Returns: path of the ansible file created
Return type: str

check_connection(rhel_pod_obj, host, pem_dst_path)¶

Check whether newly brought up RHEL instances are accepting ssh connections

Parameters

rhel_pod_obj (Pod) – object for handling ansible pod
host (str) – Node to which we want to try ssh
pem_dst_path (str) – path to private key for ssh

create_nodes(node_conf, node_type, num_nodes)¶

create aws instances of nodes

Parameters

node_conf (dict) – of node configuration
node_type (str) – type of node to be created RHCOS/RHEL
num_nodes (int) – Number of node instances to be created

Returns

of AWSUPINode objects

Return type

list

detach_volume(volume, node=None, delete_from_backend=True)¶

Detach a volume from an EC2 instance

Parameters

volume (Volume) – The volume to delete
node (OCS) – The OCS object representing the node
delete_from_backend (bool) – True for deleting the disk from the storage backend, False otherwise

get_available_slots(existing_indexes, required_slots)¶

Get indexes which are free

Parameters

existing_indexes (list) – of integers
required_slots (int) – required number of integers

Returns

of integers (available slots)

Return type

list

get_data_volumes()¶

Get the data EBS volumes

Returns: EBS Volume instances
Return type: list

get_ec2_instances(nodes)¶

Get the EC2 instances dicts

Parameters: nodes (list) – The OCS objects of the nodes
Returns: The EC2 instances dicts (IDs and names)
Return type: dict

get_existing_indexes(index_list)¶

Extract index suffixes from index_list

Parameters: index_list (list) – of stack names in the form of ‘clustername-no$i’
Returns: sorted list of Integers
Return type: list

get_node_by_attached_volume(volume)¶

Get node OCS object of the EC2 instance that has the volume attached to

Parameters: volume (Volume) – The volume to get the EC2 according to
Returns: The OCS object of the EC2 instance
Return type: OCS

get_ready_status(node_info)¶

Get the node ‘Ready’ status

Parameters: node_info (dict) – Node info which includes details
Returns: True if node is Ready else False
Return type: bool

get_stack_name_of_node(node_name)¶

Get the stack name of a given node

Parameters: node_name (str) – the name of the node
Returns: The stack name of the given node
Return type: str

restart_nodes(nodes, timeout=300, wait=True)¶

Restart EC2 instances

Parameters

nodes (list) – The OCS objects of the nodes
wait (bool) – True if need to wait till the restarted node reaches READY state. False otherwise
timeout (int) – time in seconds to wait for node to reach ‘not ready’ state, and ‘ready’ state.

restart_nodes_by_stop_and_start(nodes, wait=True, force=True)¶

Restart nodes by stopping and starting EC2 instances

Parameters

nodes (list) – The OCS objects of the nodes
wait (bool) – True in case wait for status is needed, False otherwise
force (bool) – True for force instance stop, False otherwise

Returns:

restart_nodes_by_stop_and_start_teardown()¶: Make sure all EC2 instances are up. To be used in the test teardown

start_nodes(instances=None, nodes=None, wait=True)¶

Start EC2 instances

Parameters

nodes (list) – The OCS objects of the nodes
instances (dict) – instance-id and name dict
wait (bool) – True for waiting the instances to start, False otherwise

stop_nodes(nodes, wait=True)¶

Stop EC2 instances

Parameters

nodes (list) – The OCS objects of the nodes
wait (bool) – True for waiting the instances to stop, False otherwise

terminate_nodes(nodes, wait=True)¶

Terminate EC2 instances

Parameters

nodes (list) – The OCS objects of the nodes
wait (bool) – True for waiting the instances to terminate,
otherwise (False) –

verify_nodes_added(hosts)¶

Verify RHEL workers are added

Parameters: hosts (list) – list of aws private hostnames
Raises: FailedToAddNodeException – if node addition failed

wait_for_volume_attach(volume)¶

Wait for an EBS volume to be attached to an EC2 instance. This is used as when detaching the EBS volume from the EC2 instance, re-attach should take place automatically

Parameters

volume (Volume) – The volume to wait for to be attached

Returns

True if the volume has been attached to the: instance, False otherwise

Return type

bool

class ocs_ci.ocs.platform_nodes.AWSUPINode(node_conf, node_type)¶

Bases: ocs_ci.ocs.platform_nodes.AWSNodes

Node object representing AWS upi nodes

add_cert_to_template(worker_template_path)¶

Add cert to worker template

Parameters: worker_template_path (str) – Path where template file is located

build_stack_params(index, conf)¶

Build all the params required for a stack creation

Parameters

index (int) – An integer index for this stack
conf (dict) – Node config

Returns

of param dicts

Return type

list

gather_worker_data(suffix='no0')¶

Gather various info like vpc, iam role, subnet,security group, cluster tag from existing RHCOS workers

Parameters: suffix (str) – suffix to get resource of worker node, ‘no0’ by default

get_cert_content(worker_ignition_path)¶

Get the certificate content from worker ignition file

Parameters: worker_ignition_path (str) – Path of the worker ignition file
Returns: certificate content
Return type: formatted_cert (str)

get_kube_tag(tags)¶

Fetch kubernets.io tag from worker instance

Parameters

tags (dict) – AWS tags from existing worker

Returns

key looks like: ”kubernetes.io/cluster/<cluster-name>” and value looks like “share” OR “owned”

Return type

tuple

get_rhcos_worker_template()¶

Download template and keep it locally

Returns: local path to template file
Return type: path (str)

get_worker_resource_id(resource)¶

Get the resource ID

Parameters: resource (dict) – a dictionary of stack resource
Returns: ID of worker stack resource
Return type: str

update_template_with_cert(worker_template_path, cert)¶

Update the template file with cert provided

Parameters

worker_template_path (str) – template file path
cert (str) – Certificate body

class ocs_ci.ocs.platform_nodes.AZURENodes¶

Bases: ocs_ci.ocs.platform_nodes.NodesBase

Azure Nodes class

attach_nodes_to_cluster(node_list)¶

attach_volume(volume, node)¶

create_and_attach_nodes_to_cluster(node_conf, node_type, num_nodes)¶

Create nodes and attach them to cluster Use this function if you want to do both creation/attachment in a single call

Parameters

node_conf (dict) – of node configuration
node_type (str) – type of node to be created RHCOS/RHEL
num_nodes (int) – Number of node instances to be created

create_nodes(node_conf, node_type, num_nodes)¶

detach_volume(volume, node=None, delete_from_backend=True)¶

Detach a volume from an Azure Vm instance

Parameters

volume (Disk) – The disk object required to delete a volume
node (OCS) – The OCS object representing the node
delete_from_backend (bool) – True for deleting the disk from the storage backend, False otherwise

get_data_volumes()¶

Get the data Azure disk objects

Returns: azure disk objects
Return type: list

get_node_by_attached_volume(volume)¶

Get node OCS object of the Azure vm instance that has the volume attached to

Parameters: volume (Disk) – The disk object to get the Azure Vm according to
Returns: The OCS object of the Azure Vm instance
Return type: OCS

restart_nodes(nodes, timeout=900, wait=True)¶

Restart Azure vm instances

Parameters

nodes (list) – The OCS objects of the nodes / Azure Vm instance
wait (bool) – True if need to wait till the restarted node reaches READY state. False otherwise
timeout (int) – time in seconds to wait for node to reach ‘not ready’ state, and ‘ready’ state.

restart_nodes_by_stop_and_start(nodes, timeout=540, wait=True, force=True)¶

Restart Azure vm instances by stop and start

Parameters

nodes (list) – The OCS objects of the nodes / Azure Vm instance
wait (bool) – True if need to wait till the restarted node reaches READY state. False otherwise
timeout (int) – time in seconds to wait for node to reach ‘not ready’ state, and ‘ready’ state.
force (bool) – True for force VM stop, False otherwise

restart_nodes_by_stop_and_start_teardown()¶: Make sure all VM instances up by the end of the test

start_nodes(nodes, timeout=540, wait=True)¶

Start Azure vm instances

Parameters

nodes (list) – The OCS objects of the nodes
wait (bool) – True for waiting the instances to start, False otherwise
timeout (int) – time in seconds to wait for node to reach ‘ready’ state.

stop_nodes(nodes, timeout=540, wait=True, force=True)¶

Stop Azure vm instances

Parameters

nodes (list) – The OCS objects of the nodes
wait (bool) – True for waiting the instances to stop, False otherwise
timeout (int) – time in seconds to wait for node to reach ‘not ready’ state.
force (bool) – True for force VM stop, False otherwise

wait_for_volume_attach(volume)¶

Wait for a Disk to be attached to an Azure Vm instance. This is used as when detaching the Disk from the Azure Vm instance, re-attach should take place automatically

Parameters

volume (Disk) – The Disk to wait for to be attached

Returns

True if the volume has been attached to the: instance, False otherwise

Return type

bool

class ocs_ci.ocs.platform_nodes.BaremetalNodes¶

Bases: ocs_ci.ocs.platform_nodes.NodesBase

Baremetal Nodes class

attach_nodes_to_cluster(node_list)¶

attach_volume(volume, node)¶

create_and_attach_nodes_to_cluster(node_conf, node_type, num_nodes)¶

Create nodes and attach them to cluster Use this function if you want to do both creation/attachment in a single call

Parameters

node_conf (dict) – of node configuration
node_type (str) – type of node to be created RHCOS/RHEL
num_nodes (int) – Number of node instances to be created

create_nodes(node_conf, node_type, num_nodes)¶

detach_volume(volume, node=None, delete_from_backend=True)¶

get_data_volumes()¶

get_node_by_attached_volume(volume)¶

read_default_config(default_config_path)¶

Commonly used function to read default config

Parameters: default_config_path (str) – Path to default config file
Returns: of default config loaded
Return type: dict

restart_nodes(nodes, force=True)¶

Restart Baremetal Machine

Parameters

nodes (list) – The OCS objects of the nodes
force (bool) – True for force BM stop, False otherwise

restart_nodes_teardown()¶: Make sure all BMs are up by the end of the test

start_nodes(nodes, wait=True)¶

Start Baremetal Machine

Parameters

nodes (list) – The OCS objects of the nodes
wait (bool) – Wait for node status

stop_nodes(nodes, force=True)¶

Stop Baremetal Machine

Parameters

nodes (list) – The OCS objects of the nodes
force (bool) – True for force nodes stop, False otherwise

wait_for_volume_attach(volume)¶

class ocs_ci.ocs.platform_nodes.IBMPowerNodes¶

Bases: ocs_ci.ocs.platform_nodes.NodesBase

IBM Power Nodes class

restart_nodes(nodes, timeout=540, wait=True, force=True)¶

Restart PowerNode

Parameters

nodes (list) – The OCS objects of the nodes
timeout (int) – time in seconds to wait for node to reach ‘not ready’ state, and ‘ready’ state.
wait (bool) – True if need to wait till the restarted node reaches timeout
force (bool) – True for force BM stop, False otherwise

restart_nodes_by_stop_and_start(nodes, force=True)¶

Restart PowerNodes with stop and start

Parameters

nodes (list) – The OCS objects of the nodes
force (bool) – True for force node stop, False otherwise

restart_nodes_by_stop_and_start_teardown()¶: Make sure all PowerNodes are up by the end of the test

start_nodes(nodes, force=True)¶

Start PowerNode

Parameters

nodes (list) – The OCS objects of the nodes
wait (bool) – Wait for node status

stop_nodes(nodes, force=True)¶

Stop PowerNode

Parameters

nodes (list) – The OCS objects of the nodes
force (bool) – True for force nodes stop, False otherwise

class ocs_ci.ocs.platform_nodes.NodesBase¶

Bases: object

A base class for nodes related operations. Should be inherited by specific platform classes

attach_nodes_to_cluster(node_list)¶

attach_volume(volume, node)¶

create_and_attach_nodes_to_cluster(node_conf, node_type, num_nodes)¶

Create nodes and attach them to cluster Use this function if you want to do both creation/attachment in a single call

Parameters

node_conf (dict) – of node configuration
node_type (str) – type of node to be created RHCOS/RHEL
num_nodes (int) – Number of node instances to be created

create_nodes(node_conf, node_type, num_nodes)¶

detach_volume(volume, node=None, delete_from_backend=True)¶

get_data_volumes()¶

get_node_by_attached_volume(volume)¶

read_default_config(default_config_path)¶

Commonly used function to read default config

Parameters: default_config_path (str) – Path to default config file
Returns: of default config loaded
Return type: dict

restart_nodes(nodes, wait=True)¶

restart_nodes_by_stop_and_start(nodes, force=True)¶

restart_nodes_by_stop_and_start_teardown()¶

start_nodes(nodes)¶

stop_nodes(nodes)¶

wait_for_volume_attach(volume)¶

class ocs_ci.ocs.platform_nodes.PlatformNodesFactory¶

Bases: object

A factory class to get specific nodes platform object

get_nodes_platform()¶

class ocs_ci.ocs.platform_nodes.VMWareLSONodes¶

Bases: ocs_ci.ocs.platform_nodes.VMWareNodes

VMWare LSO nodes class

detach_volume(volume, node=None, delete_from_backend=True)¶

Detach disk from a VM and delete from datastore if specified

Parameters

volume (str) – Volume path
node (OCS) – The OCS object representing the node
delete_from_backend (bool) – True for deleting the disk (vmdk) from backend datastore, False otherwise

get_data_volumes(pvs=None)¶

Get the data vSphere volumes

Parameters: pvs (list) – PV OCS objects
Returns: vSphere volumes
Return type: list

class ocs_ci.ocs.platform_nodes.VMWareNodes¶

Bases: ocs_ci.ocs.platform_nodes.NodesBase

VMWare nodes class

attach_volume(node, volume)¶

create_and_attach_nodes_to_cluster(node_conf, node_type, num_nodes)¶

Create nodes and attach them to cluster Use this function if you want to do both creation/attachment in a single call

Parameters

node_conf (dict) – of node configuration
node_type (str) – type of node to be created RHCOS/RHEL
num_nodes (int) – Number of node instances to be created

create_and_attach_volume(node, size)¶

Create a new volume and attach it to the given VM

Parameters

node (OCS) – The OCS object representing the node
size (int) – The size in GB for the new volume

detach_volume(volume, node=None, delete_from_backend=True)¶

Detach disk from a VM and delete from datastore if specified

Parameters

volume (str) – Volume path
node (OCS) – The OCS object representing the node
delete_from_backend (bool) – True for deleting the disk (vmdk) from backend datastore, False otherwise

get_data_volumes(pvs=None)¶

Get the data vSphere volumes

Parameters: pvs (list) – PV OCS objects
Returns: vSphere volumes
Return type: list

get_node_by_attached_volume(volume)¶

get_vms(nodes)¶

Get vSphere vm objects list

Parameters: nodes (list) – The OCS objects of the nodes
Returns: vSphere vm objects list
Return type: list

restart_nodes(nodes, force=True, timeout=300, wait=True)¶

Restart vSphere VMs

Parameters

nodes (list) – The OCS objects of the nodes
force (bool) – True for Hard reboot, False for Soft reboot
timeout (int) – time in seconds to wait for node to reach READY state
wait (bool) – True if need to wait till the restarted OCP node reaches READY state. False otherwise

restart_nodes_by_stop_and_start(nodes, force=True)¶

Restart vSphere VMs

Parameters

nodes (list) – The OCS objects of the nodes
force (bool) – True for force VM stop, False otherwise

restart_nodes_by_stop_and_start_teardown()¶: Make sure all VMs are up by the end of the test

start_nodes(nodes, wait=True)¶

Start vSphere VMs

Parameters: nodes (list) – The OCS objects of the nodes

stop_nodes(nodes, force=True)¶

Stop vSphere VMs

Parameters

nodes (list) – The OCS objects of the nodes
force (bool) – True for force VM stop, False otherwise

wait_for_volume_attach(volume)¶

class ocs_ci.ocs.platform_nodes.VSPHEREUPINode(node_conf, node_type, compute_count)¶

Bases: ocs_ci.ocs.platform_nodes.VMWareNodes

Node object representing VMWARE UPI nodes

add_node(use_terraform=True)¶

Add nodes to the current cluster

Parameters: use_terraform (bool) – if True use terraform to add nodes, otherwise use manual steps to add nodes

add_nodes_with_terraform()¶: Add nodes using terraform

add_nodes_without_terraform()¶: Add nodes without terraform

generate_node_names_for_vsphere(count, prefix='compute-')¶

Generate the node names for vsphere platform

Parameters

count (int) – Number of node names to generate
prefix (str) – Prefix for node name

Returns

List of node names

Return type

list

wait_for_connection_and_set_host_name(ip, host_name)¶

Waits for connection to establish to the node and sets the hostname

Parameters

ip (str) – IP of the node
host_name (str) – Host name to set for the node

Raises

NoValidConnectionsError – Raises if connection is not established
AuthenticationException – Raises if credentials are not correct

ocs_ci.ocs.pod_exec module¶

Dispatcher for proper exec based on api-client config

This class takes care of dispatching a proper api-client instance to take care of command execution.

WHY this class ? Execution of command might happen using various means, as of now we are using kubernetes client but we might also use openshift rest client. To cope with this we might have to dynamically figure out at run time which backend to use.

This module will have all api client class definitions in this file along with dispatcher class Exec.

class ocs_ci.ocs.pod_exec.CmdObj(cmd, timeout, wait, check_ec, long_running)¶

Bases: tuple

property check_ec¶: Alias for field number 3

property cmd¶: Alias for field number 0

property long_running¶: Alias for field number 4

property timeout¶: Alias for field number 1

property wait¶: Alias for field number 2

class ocs_ci.ocs.pod_exec.Exec(oc_client='KubClient')¶

Bases: object

Dispatcher class for proper api client instantiation

This class has a factory function which returns proper api client instance necessary for command execution

run(podname, namespace, cmd_obj)¶: actual run happens here Command will be fwd to api client object and it should return a 3 tuple (stdout, stderr, retval)

class ocs_ci.ocs.pod_exec.KubClient¶

Bases: object

Specific to upstream Kubernetes client library

run(podname, namespace, cmd_obj)¶

ocs_ci.ocs.pod_exec.register_class(cls)¶

Decorator for registering a class ‘cls’ in class map

Please make sure that api-client name in config should be same as class name.

ocs_ci.ocs.rados_utils module¶

class ocs_ci.ocs.rados_utils.RadosHelper(mon, config=None, log=None, cluster='ceph')¶

Bases: object

create_pool(pool_name, pg_num=16, erasure_code_profile_name=None, min_size=None, erasure_code_use_overwrites=False)¶

Create a pool named from the pool_name parameter.

Parameters

pool_name (str) – name of the pool being created.
pg_num (int) – initial number of pgs.
erasure_code_profile_name (str) – if set and !None create an erasure coded pool using the profile
min_size (int) – minimum size
erasure_code_use_overwrites (bool) – if True, allow overwrites

get_mgr_proxy_container(node, docker_image, proxy_container='mgr_proxy')¶

Returns mgr dummy container to access containerized storage :param node: ceph node :type node: ceph.ceph.CephNode :param docker_image: repository/image:tag :type docker_image: str

Returns: mgr object
Return type: ceph.ceph.CephDemon

get_num_pools()¶

Returns: number of pools in the cluster

get_osd_dump_json()¶: osd dump –format=json converted to a python object :returns: the python object

get_pg_primary(pool, pgnum)¶: get primary for pool, pgnum (e.g. (data, 0)->0

get_pg_random(pool, pgnum)¶: get random osd for pool, pgnum (e.g. (data, 0)->0

get_pgid(pool, pgnum)¶

Parameters

pool – pool name
pgnum – pg number

Returns

a string representing this pg.

get_pool_dump(pool)¶: get the osd dump part of a pool

get_pool_num(pool)¶: get number for pool (e.g., data -> 2)

get_pool_property(pool_name, prop)¶

Parameters

pool_name – pool
prop – property to be checked.

Returns

property as an int value.

is_up(osd_id)¶: :return 1 if up, 0 if down

kill_osd(osd_node, osd_service)¶

Params: id , type of signal, list of osd objects type: “SIGKILL”, “SIGTERM”, “SIGHUP” etc.
Returns: 1 or 0

list_pools()¶: list all pool names

raw_cluster_cmd(*args)¶

Returns: (stdout, stderr)

revive_osd(osd_node, osd_service)¶

Returns: 0 if revive success,1 if fail

ocs_ci.ocs.registry module¶

ocs_ci.ocs.registry.add_role_to_user(role_type, user, cluster_role=False, namespace=None)¶

Function to add a cluster/regular role to user

Parameters

role_type (str) – Type of the role to be added
user (str) – User to be added for the role
cluster_role (bool) – Whether to add a cluster-role or a regular role
namespace (str) – Namespace to be used

Raises

AssertionError – When failure in adding new role to user

ocs_ci.ocs.registry.change_registry_backend_to_ocs()¶

Function to deploy registry with OCS backend.

Raises: AssertionError – When failure in change of registry backend to OCS

ocs_ci.ocs.registry.check_image_exists_in_registry(image_url)¶

Function to check either image exists in registry or not

Parameters: image_url (str) – Image url to be verified
Returns: True if image exists, else False
Return type: bool

ocs_ci.ocs.registry.enable_route_and_create_ca_for_registry_access()¶

Function to enable route and to create ca, copy to respective location for registry access

Raises: AssertionError – When failure in enabling registry default route

ocs_ci.ocs.registry.get_build_name_by_pattern(pattern='', namespace=None)¶

In a given namespace find names of the builds that match the given pattern

Parameters

pattern (str) – name of the build with given pattern
namespace (str) – Namespace value

Returns

List of build names matching the pattern

Return type

build_list (list)

ocs_ci.ocs.registry.get_default_route_name()¶

Function to get default route name

Returns: Returns default route name
Return type: route_name (str)

ocs_ci.ocs.registry.get_oc_podman_login_cmd(skip_tls_verify=True)¶

Function to get oc and podman login commands on node

Parameters: skip_tls_verify (bool) – If true, the server’s certificate will not be checked for validity
Returns: List of cmd for oc/podman login
Return type: cmd_list (list)

ocs_ci.ocs.registry.get_registry_pod_obj()¶

Function to get registry pod obj

Returns: List of Registry pod objs
Return type: pod_obj (list)
Raises: UnexpectedBehaviour – When image-registry pod is not present.

ocs_ci.ocs.registry.image_list_all()¶

Function to list the images in the podman registry

Returns: Images present in cluster
Return type: image_list_output (str)

ocs_ci.ocs.registry.image_pull(image_url)¶

Function to pull images from repositories

Parameters: image_url (str) – Image url container image repo link

ocs_ci.ocs.registry.image_pull_and_push(project_name, template, image='', pattern='', wait=True)¶: Pull and push images running oc new-app command :param project_name: Name of project :type project_name: str :param template: Name of the template of the image :type template: str :param image: Name of the image with tag :type image: str :param pattern: name of the build with given pattern :type pattern: str :param wait: If true waits till the image pull and push completes. :type wait: bool

ocs_ci.ocs.registry.image_push(image_url, namespace)¶

Function to push images to destination

Parameters

image_url (str) – Image url container image repo link
namespace (str) – Image to be uploaded namespace

Returns

Uploaded image path

Return type

registry_path (str)

ocs_ci.ocs.registry.image_rm(registry_path, image_url)¶

Function to remove images from registry

Parameters

registry_path (str) – Image registry path
image_url (str) – Image url container image repo link

ocs_ci.ocs.registry.modify_registry_pod_count(count)¶

Function to modify registry replica count(increase/decrease pod count)

Parameters: count (int) – registry replica count to be changed to
Returns: True in case if changes are applied. False otherwise
Return type: bool
Raises: TimeoutExpiredError – When number of image registry pods doesn’t match the count

ocs_ci.ocs.registry.remove_role_from_user(role_type, user, cluster_role=False, namespace=None)¶

Function to remove a cluster/regular role from a user

Parameters

role_type (str) – Type of the role to be removed
user (str) – User of the role
cluster_role (bool) – Whether to remove a cluster-role or a regular role
namespace (str) – Namespace to be used

Raises

AssertionError – When failure in removing role from user

ocs_ci.ocs.registry.validate_image_exists(namespace=None)¶

Validate image exists on registries path :param namespace: Namespace where the images/builds are created :type namespace: str

Returns: Dir/Files/Images are listed in string format
Return type: image_list (str)
Raises: Exceptions if dir/folders not found –

ocs_ci.ocs.registry.validate_pvc_mount_on_registry_pod()¶

Function to validate pvc mounted on the registry pod

Raises: AssertionError – When PVC mount not present in the registry pod

ocs_ci.ocs.registry.validate_registry_pod_status()¶: Function to validate registry pod status

ocs_ci.ocs.ripsaw module¶

RipSaw Class to run various workloads and scale tests

class ocs_ci.ocs.ripsaw.RipSaw(**kwargs)¶

Bases: object

Workload operation using RipSaw

apply_crd(crd)¶

Apply the CRD

Parameters: crd (str) – Name of file to apply

cleanup()¶

get_uuid(benchmark)¶

Getting the UUID of the test.: when ripsaw used for running a benchmark tests, each run get its own UUID, so the results in the elastic-search server can be sorted.

Parameters: benchmark (str) – the name of the main pod in the test
Returns: the UUID of the test
Return type: str

ocs_ci.ocs.scale_lib module¶

class ocs_ci.ocs.scale_lib.FioPodScale(kind='deploymentconfig', pod_dict_path='/home/docs/checkouts/readthedocs.org/user_builds/ocs-ci-pbalogh/checkouts/stable/ocs_ci/templates/app-pods/fedora_dc.yaml', node_selector={'scale-label': 'app-scale'})¶

Bases: object

FioPodScale Class with required scale library functions and params

cleanup()¶: Function to tear down

create_and_set_namespace()¶: Create and set namespace for the pods to be created Create sa_name if Kind if DeploymentConfig

create_multi_pvc_pod(pods_per_iter=5, io_runtime=3600, start_io=False, pvc_size=None)¶

Function to create PVC of different type and attach them to PODs and start IO.

Parameters

pods_per_iter (int) – Number of PVC-POD to be created per PVC type
If 2 then 8 PVC+POD will be created with 2 each of 4 PVC types (Example,) –
io_runtime (sec) – Fio run time in seconds
start_io (bool) – If True start IO else don’t
pvc_size (Gi) – size of PVC

Returns

Objs of all the PODs created pvc_objs (obj): Objs of all the PVCs created

Return type

pod_objs (obj)

create_scale_pods(scale_count=1500, pods_per_iter=5, io_runtime=None, pvc_size=None, start_io=None)¶

Main Function with scale pod creation flow and checks to add nodes. For other platforms will not be considering the instance_type param

Parameters

scale_count (int) – Scale pod+pvc count
io_runtime (sec) – Fio run time in seconds
start_io (bool) – If True start IO else don’t
pods_per_iter (int) – Number of PVC-POD to be created per PVC type
pvc_size (Gi) – size of PVC
If 5 then 20 PVC+POD will be created with 5 each of 4 PVC types (Example,) –
value in-between 5-10 (Test) –

property kind¶

property node_selector¶

property pod_dict_path¶

pvc_expansion(pvc_new_size)¶: Function to expand PVC size and verify the new size is reflected.

ocs_ci.ocs.scale_lib.add_worker_based_on_cpu_utilization(node_count, expected_percent, role_type=None, machineset_name=None)¶

Function to evaluate CPU utilization of nodes and add node if required.

Parameters

machineset_name (list) – Machineset_names to add more nodes if required.
node_count (int) – Additional nodes to be added
expected_percent (int) – Expected utilization precent
role_type (str) – To add type to the nodes getting added

Returns

True if Nodes gets added, else false.

Return type

bool

ocs_ci.ocs.scale_lib.add_worker_based_on_pods_count_per_node(node_count, expected_count, role_type=None, machineset_name=None)¶

Function to evaluate number of pods up in node and add new node accordingly.

Parameters

machineset_name (list) – Machineset_names to add more nodes if required.
node_count (int) – Additional nodes to be added
expected_count (int) – Expected pod count in one node
role_type (str) – To add type to the nodes getting added

Returns

True if Nodes gets added, else false.

Return type

bool

ocs_ci.ocs.scale_lib.check_all_pvc_reached_bound_state_in_kube_job(kube_job_obj, namespace, no_of_pvc, timeout=30)¶

Function to check either bulk created PVCs reached Bound state using kube_job

Parameters

kube_job_obj (obj) – Kube Job Object
namespace (str) – Namespace of PVC’s created
no_of_pvc (int) – Bulk PVC count
timeout – a timeout for all the pvc in kube job to reach bound status

Returns

List of all PVCs which is in Bound state.

Return type

pvc_bound_list (list)

Asserts:: If not all PVC reached to Bound state.

ocs_ci.ocs.scale_lib.check_and_add_enough_worker(worker_count)¶

Function to check if there is enough workers available to scale pods. IF there is no enough worker then worker will be added based on supported platforms Function also adds scale label to the respective worker nodes.

Parameters: worker_count (int) – Expected worker count to be present in the setup
Returns: True is there is enough worker count else raise exception.
Return type: book

ocs_ci.ocs.scale_lib.check_enough_resource_available_in_workers(ms_name=None, pod_dict_path=None)¶

Function to check if there is enough resource in worker, if not add worker for automation supported platforms

Parameters

ms_name (list) – Require param in-case of aws platform to increase the worker
pod_dict_path (str) – Pod dict path for nginx pod.

ocs_ci.ocs.scale_lib.construct_pvc_clone_yaml_bulk_for_kube_job(pvc_dict_list, clone_yaml, sc_name)¶

Function to construct pvc.yaml to create bulk of pvc clones using kube_job

Parameters

pvc_dict_list (list) – List of PVCs for each of them one clone is to be created
clone_yaml (str) – Clone yaml which is the template for building clones
sc_name (str) – SC name for pvc creation

Returns

List of all PVC.yaml dicts

Return type

pvc_dict_list (list)

ocs_ci.ocs.scale_lib.construct_pvc_creation_yaml_bulk_for_kube_job(no_of_pvc, access_mode, sc_name, pvc_size=None)¶

Function to construct pvc.yaml to create bulk of pvc’s using kube_job

Parameters

no_of_pvc (int) – Bulk PVC count
access_mode (str) – PVC access_mode
sc_name (str) – SC name for pvc creation
pvc_size (str) – size of all pvcs to be created with Gi suffix (e.g. 10Gi). If None, random size pvc will be created

Returns

List of all PVC.yaml dicts

Return type

pvc_dict_list (list)

ocs_ci.ocs.scale_lib.delete_objs_parallel(obj_list, namespace, kind)¶

Function to delete objs specified in list

Parameters

obj_list (list) – List can be obj of pod, pvc, etc
namespace (str) – Namespace where the obj belongs to
kind (str) – Obj Kind

ocs_ci.ocs.scale_lib.get_expected_worker_count(scale_count=1500)¶

Function to get expected worker count based on platform to scale pods in cluster

Parameters: scale_count (int) – Scale count of the PVC+POD to be created
Returns: Expected worker count to scale required number of pod
Return type: expected_worker_count (int)

ocs_ci.ocs.scale_lib.get_max_pvc_count()¶: Return the maximum number of pvcs to test for. This value is 500 times the number of worker nodes.

ocs_ci.ocs.scale_lib.get_rate_based_on_cls_iops(custom_iops_dict=None, osd_size=2048)¶

Function to check ceph cluster iops and suggest rate param for fio.

Parameters

osd_size (int) – Size of the OSD in GB
custom_iops_dict (dict) – Dictionary of rate param to be used during IO run.
iops_dict = {'usage_below_40%' (Example,) – ‘16k’, ‘usage_40%_60%’: ‘8k’,
'usage_60%_80%' – ‘4k’, ‘usage_80%_95%’: ‘2K’}
Make sure dict key is same as above example. (Warning,) –

Returns

Rate parm for fio based on ceph cluster IOPs

Return type

rate_param (str)

ocs_ci.ocs.scale_lib.get_size_based_on_cls_usage(custom_size_dict=None)¶

Function to check cls capacity suggest IO write to cluster

Parameters

custom_size_dict (dict) – Dictionary of size param to be used during IO run.
size_dict = {'usage_below_60' (Example,) – ‘2G’, ‘usage_60_70’: ‘512M’,
'usage_70_80' – ‘10M’, ‘usage_80_85’: ‘512K’, ‘usage_above_85’: ‘10K’}
Make sure dict key is same as above example. (Warning,) –

Returns

IO size to be considered for cluster env

Return type

size (str)

ocs_ci.ocs.scale_lib.increase_pods_per_worker_node_count(pods_per_node=500, pods_per_core=10)¶

Function to increase pods per node count, default OCP supports 250 pods per node, from OCP 4.6 limit is going to be 500, but using this function can override this param to create more pods per worker nodes. more detail: https://docs.openshift.com/container-platform/4.5/nodes/nodes/nodes-nodes-managing-max-pods.html

Example: The default value for podsPerCore is 10 and the default value for maxPods is 250. This means that unless the node has 25 cores or more, by default, podsPerCore will be the limiting factor.

WARN: This function will perform Unscheduling of workers and reboot so Please aware if there is any non-dc pods then expected to be terminated.

Parameters

pods_per_node (int) – Pods per node limit count
pods_per_core (int) – Pods per core limit count

Raise:: UnexpectedBehaviour if machineconfigpool not in Updating state within 40secs.

ocs_ci.ocs.scale_pgsql module¶

ScalePodPGSQL workload class for scale

class ocs_ci.ocs.scale_pgsql.ScalePodPGSQL(node_selector={'scale-label': 'app-scale'}, **kwargs)¶

Bases: ocs_ci.ocs.pgsql.Postgresql

Scale Postgresql workload with scale parameters and functions

apply_crd(crd)¶

Apply the CRD

Parameters: crd (str) – Name of file to apply

cleanup()¶: Clean up

setup_postgresql(replicas, node_selector=None)¶

Deploy postgres sql server

Parameters: replicas (int) – Number of postgresql pods to be deployed
Raises: CommandFailed – If PostgreSQL server setup fails

ocs_ci.ocs.scale_pgsql.add_worker_node(instance_type=None)¶

ocs_ci.ocs.scale_pgsql.delete_worker_node()¶

ocs_ci.ocs.small_file_workload module¶

Test to exercise Small File Workload

Note: This test is using the ripsaw and the elastic search, so it start process with port forwarding on port 9200 from the host that run the test (localhost) to the elastic-search within the open-shift cluster, so, if you host is listen to port 9200, this test can not be running in your host.

ocs_ci.ocs.small_file_workload.smallfile_workload(ripsaw, es, file_size, files, threads, samples, interface)¶

Run SmallFile Workload SmallFile workload uses https://github.com/distributed-system-analysis/smallfile

smallfile is a python-based distributed POSIX workload generator which can be used to quickly measure performance and scaling for a variety of metadata-intensive workloads

Parameters

-- Ripsaw fixture to setup/teardown ripsaw (ripsaw) –
-- Elastic search fixture (es) –
-- size of file to be created (file_siles) –
-- number of files to be created (files) –
-- number of threads to run (threads) –
-- samples taken if running performance tests (samples) –
-- CephFileSystem or CephBlockPool (interface) –

Returns

backup_es (str) – backup elastic search location environment (dict) – environment (user is changed at times) sf_data (dict) – small file data extracted from yaml and modified by this routine

ocs_ci.ocs.uninstall module¶

ocs_ci.ocs.uninstall.remove_cluster_logging_operator_from_ocs()¶: Function removes cluster logging operator from OCS

ocs_ci.ocs.uninstall.remove_monitoring_stack_from_ocs()¶: Function removes monitoring stack from OCS

ocs_ci.ocs.uninstall.remove_ocp_registry_from_ocs(platform)¶

Function removes OCS registry from OCP cluster

Parameters: platform (str) – the platform the cluster deployed on

ocs_ci.ocs.uninstall.uninstall_lso(lso_sc)¶: Function uninstalls local-volume objects from OCS cluster

ocs_ci.ocs.uninstall.uninstall_ocs()¶: The function uninstalls the OCS operator from a openshift cluster and removes all its settings and dependencies

ocs_ci.ocs.utils module¶

ocs_ci.ocs.utils.apply_oc_resource(template_name, cluster_path, _templating, template_data=None, template_dir='ocs-deployment')¶

Apply an oc resource after rendering the specified template with the rook data from cluster_conf.

Parameters

template_name (str) – Name of the ocs-deployment config template
cluster_path (str) – Path to cluster directory, where files will be written
_templating (Templating) – Object of Templating class used for templating
template_data (dict) – Data for render template (default: {})
template_dir (str) – Directory under templates dir where template exists (default: ocs-deployment)

ocs_ci.ocs.utils.check_ceph_healthly(ceph_mon, num_osds, num_mons, mon_container=None, timeout=300)¶

Function to check ceph is in healthy state

Parameters

ceph_mon (CephNode) – monitor node
num_osds (int) – number of osds in cluster
num_mons (int) – number of mons in cluster
mon_container (str) – monitor container name if monitor is placed in the container
timeout – 300 seconds(default) max time to check if cluster is not healthy within timeout period return 1

Returns

returns 0 when ceph is in healthy state otherwise returns 1

Return type

int

ocs_ci.ocs.utils.cleanup_ceph_nodes(osp_cred, pattern=None, timeout=300)¶

ocs_ci.ocs.utils.collect_noobaa_db_dump(log_dir_path)¶

Collect the Noobaa DB dump

Parameters: log_dir_path (str) – directory for dumped Noobaa DB

ocs_ci.ocs.utils.collect_ocs_logs(dir_name, ocp=True, ocs=True, mcg=False, status_failure=True)¶

Collects OCS logs

Parameters

dir_name (str) – directory name to store OCS logs. Logs will be stored in dir_name suffix with _ocs_logs.
ocp (bool) – Whether to gather OCP logs
ocs (bool) – Whether to gather OCS logs
mcg (bool) – True for collecting MCG logs (noobaa db dump)
status_failure (bool) – Whether the collection is after success or failure, allows better naming for folders under logs directory

ocs_ci.ocs.utils.collect_prometheus_metrics(metrics, dir_name, start, stop, step=1.0)¶

Collects metrics from Prometheus and saves them in file in json format. Metrics can be found in OCP Console in Monitoring -> Metrics.

Parameters

metrics (list) – list of metrics to get from Prometheus (E.g. ceph_cluster_total_used_bytes, cluster:cpu_usage_cores:sum, cluster:memory_usage_bytes:sum)
dir_name (str) – directory name to store metrics. Metrics will be stored in dir_name suffix with _ocs_metrics.
start (str) – start timestamp of required datapoints
stop (str) – stop timestamp of required datapoints
step (float) – step of required datapoints

ocs_ci.ocs.utils.config_ntp(ceph_node)¶

ocs_ci.ocs.utils.create_ceph_conf(fsid, mon_hosts, pg_num='128', pgp_num='128', size='2', auth='cephx', pnetwork='172.16.0.0/12', jsize='1024')¶

ocs_ci.ocs.utils.create_ceph_nodes(cluster_conf, inventory, osp_cred, run_id, instances_name=None)¶

ocs_ci.ocs.utils.create_nodes(conf, inventory, osp_cred, run_id, instances_name=None)¶

ocs_ci.ocs.utils.create_oc_resource(template_name, cluster_path, _templating, template_data=None, template_dir='ocs-deployment')¶

Create an oc resource after rendering the specified template with the rook data from cluster_conf.

Parameters

template_name (str) – Name of the ocs-deployment config template
cluster_path (str) – Path to cluster directory, where files will be written
_templating (Templating) – Object of Templating class used for templating
template_data (dict) – Data for render template (default: {})
template_dir (str) – Directory under templates dir where template exists (default: ocs-deployment)

ocs_ci.ocs.utils.generate_repo_file(base_url, repos)¶

ocs_ci.ocs.utils.get_ceph_versions(ceph_nodes, containerized=False)¶

Log and return the ceph or ceph-ansible versions for each node in the cluster.

Parameters

ceph_nodes – nodes in the cluster
containerized – is the cluster containerized or not

Returns

A dict of the name / version pair for each node or container in the cluster

ocs_ci.ocs.utils.get_cluster_object(external_rhcs_info)¶

Build a external_ceph.ceph object with all node and role info

Parameters: external_rhcs_info (dict) –
Returns: external_ceph.ceph object

ocs_ci.ocs.utils.get_external_mode_rhcs()¶

Get external cluster info from config and obtain external cluster object

Returns: external_ceph.Ceph object

ocs_ci.ocs.utils.get_iso_file_url(base_url)¶

ocs_ci.ocs.utils.get_openstack_driver(yaml)¶

ocs_ci.ocs.utils.get_pod_name_by_pattern(pattern='client', namespace=None, filter=None)¶

In a given namespace find names of the pods that match the given pattern

Parameters

pattern (str) – name of the pod with given pattern
namespace (str) – Namespace value
filter (str) – pod name to filter from the list

Returns

List of pod names matching the pattern

Return type

pod_list (list)

ocs_ci.ocs.utils.get_public_network()¶

Get the configured public network subnet for nodes in the cluster.

Returns: (str) public network subnet

ocs_ci.ocs.utils.get_root_permissions(node, path)¶: Transfer ownership of root to current user for the path given. Recursive. :param node: :type node: ceph.ceph.CephNode :param path: file path

ocs_ci.ocs.utils.hard_reboot(gyaml, name=None)¶

ocs_ci.ocs.utils.keep_alive(ceph_nodes)¶

ocs_ci.ocs.utils.kill_osd_external(ceph_cluster, osd_id, sig_type='SIGTERM')¶

Kill an osd with given signal

Parameters

ceph_cluster (external_cluster.Ceph) – Cluster object
osd_id (int) – id of osd
sig_type (str) – type of signal to be sent

Raises

CommandFailed exception –

ocs_ci.ocs.utils.node_power_failure(gyaml, sleep_time=300, name=None)¶

ocs_ci.ocs.utils.oc_get_all_obc_names()¶

Returns: A set of all OBC names
Return type: set

ocs_ci.ocs.utils.open_firewall_port(ceph_node, port, protocol)¶: Opens firewall ports for given node :param ceph_node: ceph node :type ceph_node: ceph.ceph.CephNode :param port: port :type port: str :param protocol: protocol :type protocol: str

ocs_ci.ocs.utils.reboot_node(ceph_node, timeout=300)¶

Reboot a node with given ceph_node object

Parameters

ceph_node (CephNode) – Ceph node object representing the node.
timeout (int) – Wait time in seconds for the node to comeback.

Raises

SSHException – if not able to connect through ssh

ocs_ci.ocs.utils.revive_osd_external(ceph_cluster, osd_id)¶

Start an already stopped osd

Parameters

ceph_cluster (external_cluster.Ceph) – cluster object
osd_id (int) – id of osd

Raises

CommandFailed exception in case of failure –

ocs_ci.ocs.utils.run_must_gather(log_dir_path, image, command=None)¶

Runs the must-gather tool against the cluster

Parameters

log_dir_path (str) – directory for dumped must-gather logs
image (str) – must-gather image registry path
command (str) – optional command to execute within the must-gather image

ocs_ci.ocs.utils.search_ethernet_interface(ceph_node, ceph_node_list)¶

Search interface on the given node node which allows every node in the cluster accesible by it’s shortname.

Parameters

ceph_node (ceph.ceph.CephNode) – node where check is performed
ceph_node_list (list) – node list to check

ocs_ci.ocs.utils.set_cdn_repo(node, repos)¶

ocs_ci.ocs.utils.setup_cdn_repos(ceph_nodes, build=None)¶

ocs_ci.ocs.utils.setup_ceph_toolbox(force_setup=False)¶

Setup ceph-toolbox - also checks if toolbox exists, if it exists it behaves as noop.

Parameters: force_setup (bool) – force setup toolbox pod

ocs_ci.ocs.utils.setup_deb_cdn_repo(node, build=None)¶

ocs_ci.ocs.utils.setup_deb_repos(node, ubuntu_repo)¶

ocs_ci.ocs.utils.setup_repos(ceph, base_url, installer_url=None)¶

ocs_ci.ocs.utils.setup_vm_node(node, ceph_nodes, **params)¶

ocs_ci.ocs.utils.store_cluster_state(ceph_cluster_object, ceph_clusters_file_name)¶

ocs_ci.ocs.utils.update_ca_cert(node, cert_url, timeout=120)¶

ocs_ci.ocs.utils.write_docker_daemon_json(json_text, node)¶: Write given string to /etc/docker/daemon/daemon :param json_text: json string :param node: Ceph node object :type node: ceph.ceph.CephNode

ocs_ci.ocs.version module¶

Version reporting module for OCS QE purposes:

logging version of OCS/OCP stack for every test run

generating version report directly usable in bug reports

It asks openshift for:

ClusterVersion resource

image identifiers of all containers running in openshift storage related namespaces

rpm package versions of few selected components (such as rook or ceph, if given pod is running)

ocs_ci.ocs.version.get_environment_info()¶

Getting the environment information, Information that will be collected

Versions:: OCP - version / build / channel OCS - version / build Ceph - version Rook - version
Platform:: BM / VmWare / Cloud provider etc. Instance type / architecture Cluster name User name that run the test

Returns: dictionary that contain the environment information
Return type: dict

ocs_ci.ocs.version.get_ocs_version()¶

Query OCP to get all information about OCS version.

Returns: ClusterVersion k8s object dict: image_dict with information about images IDs
Return type: dict

ocs_ci.ocs.version.main()¶: Main fuction of version reporting command line tool. used by entry point report-version from setup.py to invoke this function.

ocs_ci.ocs.version.report_ocs_version(cluster_version, image_dict, file_obj)¶

Report OCS version via:

python logging
printing human readable version into file_obj (stdout by default)

Parameters

cluster_version (dict) – cluster version dict
image_dict (dict) – dict of image objects
file_obj (object) – file object to log information

ocs_ci.ocs.workload module¶

class ocs_ci.ocs.workload.WorkLoad(name=None, path=None, work_load=None, storage_type='fs', pod=None, jobs=1)¶

Bases: object

run(**conf)¶

Perform work_load_mod.run in order to run actual io. Every workload module should implement run() function so that we can invoke <workload_module>.run() to run IOs.

Parameters: **conf (dict) – Run configuration a.k.a parameters for workload io runs
Returns: Returns a concurrent.future object
Return type: result (Future)

setup(**setup_conf)¶

Perform work_load_mod.setup() to setup the workload. Every workload module should implement setup() method so that respective <workload_module>.setup() function can be called from here

Parameters: setup_conf (dict) – Work load setup configuration, varies from workload to workload. Refer constants.TEMPLATE_WORKLOAD_DIR for various available workloads
Returns: True if setup is success else False
Return type: bool

Module contents¶

Avoid already-imported warning cause of we are importing this package from run wrapper for loading config.

You can see documentation here: https://docs.pytest.org/en/latest/reference.html under section PYTEST_DONT_REWRITE