Ubuntu upgrade from 20.04 to 22.04
40 views
Skip to first unread message
chris stokes
Feb 28, 2025, 10:02:33 AMFeb 28
to codership
Hi,
I have upgraded a node from ubuntu 20.04 to to 22.04, mysql version is the same. after the upgrade and making sure the correct packages are installed i get this error.
2025-02-28 14:21:04 0 [Warning] WSREP: handshake with 00000000-0000 failed: 'duplicate uuid'
2025-02-28 14:21:04 0 [ERROR] WSREP: failed to open gcomm backend connection: 131: A node with the same UUID already exists in the cluster. Removing gvwstate.dat file, this node will generate a new UUID when restarted. (FATAL)
at ./gcomm/src/gmcast_proto.cpp:handle_failed():313
2025-02-28 14:21:04 0 [ERROR] WSREP: ./gcs/src/gcs_core.cpp:gcs_core_open():221: Failed to open backend connection: -131 (State not recoverable)
I have tried getting rid of /var/lib/mysql and seeing if the other 6 primary node will initiate a full SST but no luck. I know that if I bring down the cluster and issue galera_new_cluster then all the nodes will rejoin correctly but i am trying to avoid doing this. I also saw this problem on two other upgraded nodes, one was a fresh install the other and upgrade, and i thought i had worked it out as after i issued an evict on the UUID that was joining it seemed to fix the issue and the node joined correctly. I don't think that is the fix anymore though. I simply can not get this upgraded node to join the cluster. Any hints on what to do about this 'duplicate uuid' would really be appreciated.
Regards
Chris
I have upgraded a node from ubuntu 20.04 to to 22.04, mysql version is the same. after the upgrade and making sure the correct packages are installed i get this error.
2025-02-28 14:21:04 0 [Warning] WSREP: handshake with 00000000-0000 failed: 'duplicate uuid'
2025-02-28 14:21:04 0 [ERROR] WSREP: failed to open gcomm backend connection: 131: A node with the same UUID already exists in the cluster. Removing gvwstate.dat file, this node will generate a new UUID when restarted. (FATAL)
at ./gcomm/src/gmcast_proto.cpp:handle_failed():313
2025-02-28 14:21:04 0 [ERROR] WSREP: ./gcs/src/gcs_core.cpp:gcs_core_open():221: Failed to open backend connection: -131 (State not recoverable)
I have tried getting rid of /var/lib/mysql and seeing if the other 6 primary node will initiate a full SST but no luck. I know that if I bring down the cluster and issue galera_new_cluster then all the nodes will rejoin correctly but i am trying to avoid doing this. I also saw this problem on two other upgraded nodes, one was a fresh install the other and upgrade, and i thought i had worked it out as after i issued an evict on the UUID that was joining it seemed to fix the issue and the node joined correctly. I don't think that is the fix anymore though. I simply can not get this upgraded node to join the cluster. Any hints on what to do about this 'duplicate uuid' would really be appreciated.
Regards
Chris
Wayne Gemmell
Mar 3, 2025, 4:26:35 AMMar 3
to chris stokes, codership
Hi Chris
Fellow user here. I have the same problem from time to time and just solve it by retrying a few times. This may not be a useful reply but it's my experience so I'm just sharing it.
Regards
Wayne
--
You received this message because you are subscribed to the Google Groups "codership" group.
To unsubscribe from this group and stop receiving emails from it, send an email to codership-tea...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/codership-team/3092f324-603c-4232-93cc-5114ff82ece6n%40googlegroups.com.
Chris Stokes
Mar 3, 2025, 9:04:37 AMMar 3
to codership
Hey Wayne
Thanks for your response. Yes that seemed to be the case for the other two nodes but the latest one refused, I even left it for a couple of days. Yesterday I needed it back in the cluster so I ended up bootstrapping again and it joined first time. Its super annoying, I can't just take down the cluster on a whim.
Thanks again
Chris
Reply all
Reply to author
Forward
0 new messages