Description of problem: ------------------- if a volume is mounted using a server IP, and if the server is being detached from the trusted pool, it can lead to data inconsistencies as explained in BZ#1639566 To avoid this, we must throw a warning prompt, when peer detach is issued , like below: "first make sure that all clients mounted using the node being detached, are remounted using the ip of another server in the trusted pool, to avoid inconsistencies", if you have done that then only proceed{y/n} Version-Release number of selected component (if applicable): ============= 3.12.2-22
Does this happen if you use backup volfile server IPs in the client mount? If not, then I am not inclined to add this warning.
Could you explain why the severity of this BZ has been marked as urgent?
(In reply to Atin Mukherjee from comment #2) > Does this happen if you use backup volfile server IPs in the client mount? > If not, then I am not inclined to add this warning. Will try this and get back
Keeping the needinfo intact. Request the severity to be also adjusted.
Nag - We need to make a decision on this bug and hence would appreciate if the needinfo can be addressed.
I managed to test this with backup volfile server and I see the behaviour being the same. I'm sending a patch in upstream to add a warning message in the peer detach code path.
upstream patch : https://review.gluster.org/#/c/glusterfs/+/21572
fyi raised 1662511 - heketi node remove/delete succeeds but gluster peer status still shows the node in trusted pool
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:0263