Link to the Document URL : https://access.redhat.com/documentation/en-us/red_hat_openshift_container_storage/3.11/html-single/deployment_guide/#chap-Documentation-Red_Hat_Gluster_Storage_Container_Native_with_OpenShift_Platform-Upgrade Describe the Issue : I see that there is no section in the doc which talks about upgrading the glusterfs_registry pods. I only see a note which asks the customer or users to follow the steps used to upgrade glusterfs pods. But this will not be sufficient in upgrading because user will need to change the values of clustername, heketi_route, node_selector in heketi and glusterfs templates of glusterfs-registry namespace which are not same as the ones in gluster namespace. Suggestions for improvement : Add a new section on upgrading glusterfs_registry pods Section Name & Number : Chapter 6. Upgrading your Red Hat Openshift Container Storage in Converged Mode Additional info : heketi_template : ======================== - description: Set the hostname for the route URL displayName: heketi route name name: HEKETI_ROUTE value: heketi-registry - displayName: heketi container image name name: IMAGE_NAME required: true value: rhgs3/rhgs-volmanager-rhel7 - displayName: heketi container image version name: IMAGE_VERSION required: true value: v3.11 - description: A unique name to identify this heketi service, useful for running multiple heketi instances displayName: GlusterFS cluster name name: CLUSTER_NAME value: registry glusterfs template: ================================ - description: Labels which define the daemonset node selector. Must contain at least one label of the format \'glusterfs=<CLUSTER_NAME>-host\' displayName: Daemonset Node Labels name: NODE_LABELS value: '{ "glusterfs": "registry-host" }' - displayName: GlusterFS container image name name: IMAGE_NAME required: true value: rhgs3/rhgs-server-rhel7 - displayName: GlusterFS container image version name: IMAGE_VERSION required: true value: v3.11 - description: A unique name to identify which heketi service manages this cluster, useful for running multiple heketi instances displayName: GlusterFS cluster name name: CLUSTER_NAME value: registry
same holds good for 3.9 and 3.10 docs as well
We do need to have a section about upgrading glusterblockprovisioner registry dc as well when we address this bug. we need to make sure that template is edited as below: ========================================================= oc edit template glusterblock-provisioner - displayName: glusterblock provisioner container image name name: IMAGE_NAME required: true value: rhgs3/rhgs-gluster-block-prov-rhel7 - displayName: glusterblock provisioner container image version name: IMAGE_VERSION required: true value: v3.10 - description: The namespace in which these resources are being created displayName: glusterblock provisioner namespace name: NAMESPACE required: true value: glusterfs-registry - description: A unique name to identify which heketi service manages this cluster, useful for running multiple heketi instances displayName: GlusterFS cluster name name: CLUSTER_NAME value: registry we need to add a step to delete resources from old pod which are below: =================================================================== Delete the following resources from the old pod # oc delete clusterroles.authorization.openshift.io glusterblock-provisioner-runner # oc delete serviceaccounts glusterblock-registry-provisioner
Hello Bhavana, Should be good until and unless we capture all the steps required and put up there. No issues. I thought having it a different section would be good as it might make users life easy when they try to upgrade or search for this particular section . Thanks kasturi
Hello Bhavana, I reviewed the doc and below are few review comments. 1) section title has to be changed to upgrade heketi registry pod 2) step1 is not required as cns-deploy and heketi client packages are already updated during the upgrade of glusterfs_registry pods 3) point No.2 we should say as backup heketi registry database 4) In step5, for ansible replace the existing command as below # oc delete deploymentconfig,service,route heketi-registry 5) In step6, example given should be changed as below. # service "heketi-registry" created route "heketi-registry" created deploymentconfig-registry "heketi" created 6) In point no.7 example given needs to be changed as below # oc get pods NAME READY STATUS RESTARTS AGE glusterblock-registry-provisioner-dc-1-nvnhc 1/1 Running 1 7d glusterfs-registry-4cpcc 1/1 Running 0 7d glusterfs-registry-9xj78 1/1 Running 0 7d glusterfs-registry-b9p5j 1/1 Running 0 7d heketi-registry-1-rr4jk 1/1 Running 0 2m 7) Title for 6.2.3 should be changed as Upgrading the Red Hat Gluster Storage registry Pods 8) point no.1 which says "Execute the following steps to enable server.tcp-user-timeout on all volumes." from 6.2.3 needs to be removed. 9) Where ever required we should be saying glusterfs-registry pods and not gluster pods. 10) example in these sections needed to be changed as below. # oc get pods NAME READY STATUS RESTARTS AGE glusterblock-registry-provisioner-dc-1-nvnhc 1/1 Running 1 7d glusterfs-registry-4cpcc 1/1 Running 0 7d glusterfs-registry-9xj78 1/1 Running 0 7d glusterfs-registry-b9p5j 1/1 Running 0 7d heketi-registry-1-rr4jk 1/1 Running 0 2m 11) oc rsh <gluster_pod> needs to be changed to <gluster_registry> pod 12) Note in this section can be removed which talks about upgrading registry pods as this section itself talks about upgrading registry pods 13) Point no.15 should say If a gluster-block-registry-provisoner-pod exists 14) replace the command # oc delete dc <gluster-block-dc> with # oc delete dc <gluster-block-registry-dc> 15) example needs to be changed to # oc delete dc glusterblock-registry-provisioner-dc 16) please confirm if the steps for cns-deploy works for registry from dev. 17) replace the command in point no.18 from oc delete serviceaccounts glusterblock-storage-provisioner to #oc delete serviceaccounts glusterblock-registry-provisioner 18) Remove the note after step 13 in section 6.1.3 as this is not required since we are having a new section on how to upgrade registry pods 19) Remove the step 1 from section 6.1.3 which says Execute the following steps to enable server.tcp-user-timeout on all volumes. Moving this bug to assigned state due to the review comments above. Thanks kasturi
Currently the topic is: 6.2. Upgrading the heketi registry pods It should be: 6.2. Upgrading the registry pods -- <<Can you please confirm if the steps for cns-deploy for registry remains the same? If yes, I can move this bug On QA. If it is different, then can you please share the steps. Looks fine.
Changes looks good to me. Since the major part of the bug is addressed i am going to move this to verified state and raise bugs for minor changes which is required.