Description of problem: TOC-> https://docs.google.com/document/d/1SwmTffJBD-Sbxx-d8ZjOOV_4qRGX8F1X8p_c4yzfm6A/edit?usp=sharing Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Hi Bhavana, 1. Need to remove yum install cns-deploy packages everywhere from commands and the headings. I see cns-deploy package installation is still written. 2. Subscribe to the OpenShift Container Platform 3.9 repository is not required. 3. I don't think atomic host should be there. We are not qualifying for atomic hosts. 4. In section, 3.2.3. Red Hat OpenShift Container Platform and Red Hat Openshift Container Storage Requirements. We don't need the 1st two multipathing steps. It's handled by openshift-glusterfs. The next two points are the only ones which need to be done manually as of now. There is a bug with these issues in POST. Adding needinfo on Jose to confirm. 5. Within the same section, why have we added 3.9 links? The links added are for cns-deploy related installation steps. 6. Basic RHOCS installation is not written in the doc. 7. In the tabular column we haven't checked for applications in the first four rows. 8. In section, 4.3. Deploying Red Hat Openshift Container Storage in Converged Mode in point no.5, Add running config.yml playboook in the first bullet. 9. If we are writing for an initial OCP installation. We should mention the pre-requisite steps for OCP as well. 10. Point 8 mentioned for sections 4.3 have to be incorporated for all the deployment sections.
Ashmitha, could you please link the bug in question? I can't keep track of which one you're referring to. :)
The PR has been created to replace verbiage in the Deploy guide for'Run the installer.' in all sections detailed below. If 'Run the installer.' does not exist then add the verbiage in the PR + the instruction for how to run the 3 openshift-ansible playbooks. PR - https://github.com/openshift/openshift-docs/pull/11234 4.3. Deploying Red Hat Openshift Container Storage in Converged Mode 4.4. Deploying Red Hat Openshift Container Storage in Converged Mode with Registry 4.5. Deploying Red Hat Openshift Container Storage in Converged Mode with Logging and Metrics 4.6. Deploying Red Hat Openshift Container Storage in Converged mode for Applications with Registry, Logging, and Metrics 5.3. Deploying Red Hat Openshift Container Storage in Independent Mode 5.4. Deploying Red Hat Openshift Container Storage in Independent mode for Applications with Registry, Logging, and Metrics
Please send complete inventory file. Do you have this inventory file option? # this option will be used for Logging and Metrics deployment openshift_master_dynamic_provisioning_enabled=true
Created attachment 1475117 [details] Inventory File for OCP on AWS for 2xOCS + Logging and Metrics Annette's inventory file for OCP 3.10 on AWS.
In Jose inventory file please commented-out the options below given they are duplicates for options taken from the documentation and in the inventory file now (openshift_metrics_storage_volume_size=10Gi and openshift_logging_es_cluster_size=3 ARE CORRECT). - openshift_metrics_cassandra_pvc_size=10Gi - openshift_logging_es_number_of_replicas=1 + #openshift_metrics_cassandra_pvc_size=10Gi + #openshift_logging_es_number_of_replicas=1 This options can also be removed completely or commented-out, it is not needed as the default replica count for metrics=1. - openshift_metrics_cassandra_replicas=1 + #openshift_metrics_cassandra_replicas=1
*** Bug 1575828 has been marked as a duplicate of this bug. ***
Also, actual inventory files had 50Gi, not 100Gi # logging openshift_logging_install_logging=true openshift_logging_es_pvc_dynamic=true openshift_logging_es_pvc_size=50Gi openshift_logging_es_cluster_size=3 openshift_logging_es_memory_limit=4Gi openshift_logging_es_pvc_storage_class_name='glusterfs-storage-block' openshift_logging_kibana_nodeselector={"node-role.kubernetes.io/infra": "true"} openshift_logging_curator_nodeselector={"node-role.kubernetes.io/infra": "true"} openshift_logging_es_nodeselector={"node-role.kubernetes.io/infra": "true"} # Container image to use for glusterfs pods openshift_storage_glusterfs_image="brew-pulp-docker01.web.prod.ext.phx2.redhat.com:8888/rhgs3/rhgs-server-rhel7:3.4.0-6" # Container image to use for glusterblock-provisioner pod openshift_storage_glusterfs_block_image="brew-pulp-docker01.web.prod.ext.phx2.redhat.com:8888/rhgs3/rhgs-gluster-block-prov-rhel7:3.4.0-4" # Container image to use for heketi pods openshift_storage_glusterfs_heketi_image="brew-pulp-docker01.web.prod.ext.phx2.redhat.com:8888/rhgs3/rhgs-volmanager-rhel7:3.4.0-3" # OCS storage cluster openshift_storage_glusterfs_namespace=app-storage openshift_storage_glusterfs_storageclass=true openshift_storage_glusterfs_storageclass_default=false openshift_storage_glusterfs_block_deploy=true penshift_storage_glusterfs_block_host_vol_create=true openshift_storage_glusterfs_block_host_vol_size=520 openshift_storage_glusterfs_block_storageclass=true openshift_storage_glusterfs_block_storageclass_default=false
Ken the documentation found at https://access.qa.redhat.com/documentation/en-us/red_hat_openshift_container_storage/3.10/html-single/deployment_guide/#chap-Documentation-Red_Hat_Gluster_Storage_Container_Native_with_OpenShift_Platform-crs_reg_log_met (5.4. Deploying Red Hat Openshift Container Storage in Independent mode for Applications with Registry, Logging, and Metrics) is a TWO CLUSTER example. The options which you say are missing are below in the 2nd cluster definition. openshift_storage_glusterfs_registry_namespace=infra-storage openshift_storage_glusterfs_registry_storageclass=false openshift_storage_glusterfs_registry_storageclass_default=false openshift_storage_glusterfs_registry_block_deploy=true openshift_storage_glusterfs_registry_block_host_vol_create=true openshift_storage_glusterfs_registry_block_host_vol_size=100 openshift_storage_glusterfs_registry_block_storageclass=true openshift_storage_glusterfs_registry_block_storageclass_default=false ALL other options in this section should stay as they are. AND there is no need to add 'openshift_logging_es_cluster_size=3' or 'openshift_logging_es_memory_limit=4Gi'. They were in my example and I am sorry I caused confusion. The 'openshift_logging_es_memory_limit=4Gi' MAY be needed for QE testing if they have limited mem on their nodes. I believe the default Mem requirment for ElasticSearch pods is 8GB Mem so if this is not avaialble the logging deployment will fail because the ES pod(s) cannot schedule.
Hi Annette, According to our documentation, if we already have an existing OCP deployed then we'll need to run just the openshift-glusterfs/config.yml playbook. In this case (brownfield deployment), the openshift-glusterfs/config.yml playbook internally only imports the openshift-glusterfs/registry.yml playbook and does NOT internally import openshift-logging/config.yml and openshift-metrics/config.yml playbooks for logging and metrics deployment respectively. You're right in comment #49, in a fresh deployment where there is no deployment on the machines, in this case the running we run prerequisite.yml and config.yml playbooks to deploy OCP, OCS with logging,metrics and registry. This works. And running prerequisite.yml and deploy_cluster.yml is sufficient. I'm stressing on brownfield deployments and it's not a corner case! Please see point 5 in all our converged mode deployment sections - 4.2, 4.3, 4.4, 4.5 and point 4 in independent mode deployment sections - 5.3 and 5.4 It's mentioned in the 2nd bullet: "For a standalone installation onto an existing OpenShift Container Platform cluster"
Also, when we are deploying OCS with logging, metrics and registry (section 5.4 and section 4.5) then this is required because the user wants these services deployed at together.