rhel-osp-director: [Doc] Replacing Controller Nodes for 8.0 is missing. Tried to follow the steps for 7: https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux_OpenStack_Platform/7/html/Director_Installation_and_Usage/Replacing_Controller_Nodes.html The pcs cluster was in maintenance mode, so I couldn't even start the keystone. After manually disabling the maintenance mode on pcs cluster, still wasn't able to successfully complete re-deployment of the overcloud as mentioned in step 7 in the guide above.
*** Bug 1313528 has been marked as a duplicate of this bug. ***
New link to this content for 7.0: https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux_OpenStack_Platform/7/html/Director_Installation_and_Usage/sect-Scaling_the_Overcloud.html#Replacing_Controller_Nodes
Sorry this took so long. I've only just been able to successfully deploy a HA Overcloud on OSP 8. Sasha, I ran into the same error you did but I think I managed to figure it out. Essentially, there's still some rouge entries for overcloud-controller-1 still in Pacemaker/Corosync, which is causing pcs to auth in a loop. However, once I cleared them away, the deployment seems to be progressing fine. I'll update the procedure tomorrow and that way you should be to test it out.
Okay, tested it out. It seems to be the failed node entry in corosync.conf that is stopping the deployment from continuing. I've got that step closer to the end, but I'll bump it up to before step 6 (restarting pacemaker and corosync). Based on my testing, this should work.
I think this might be a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=1327701 In any case, I'll switch this over the ON_QA too. As mentioned int he other BZ, here's the latest draft: https://access.stage.redhat.com/documentation/en/red-hat-openstack-platform/8/director-installation-and-usage/94-replacing-controller-nodes Sasha, is there anything else needed for this procedure?
Verified: This section of the doc looks good.
Changes now live on the customer portal.