[Gluster-users] Growing cluster: peering worked, staging failed

[Andreas Schwibbe]

Fellow gluster users,

trying to extend a 3 node cluster that is serving me very reliably for
a long time now.
Cluster is serving two volumes:

Volume Name: gv0
Type: Distributed-Replicate
Volume ID: 9bafc4d2-d9b6-4b6d-a631-1cf42d1d2559
Status: Started
Snapshot Count: 0
Number of Bricks: 6 x (2 + 1) = 18
Transport-type: tcp

Volume Name: gv1
Type: Replicate
Volume ID: 69a12600-6720-4e96-a269-931d72d4953e
Status: Started
Snapshot Count: 0
Number of Bricks: 1 x 3 = 3
Transport-type: tcp

adding peer to the cluster with
gluster peer probe node4 

works. 
gluster peer status & gluster pool list 

show all peers connected on every node.

However,
when doing
gluster v status

causes:
Staging failed on node4. Error: Volume gv0 does not exist
Staging failed on node4. Error: Volume gv1 does not exist

on node4 I can see that /var/lib/glusterd/vols is empty!

/var/log/glusterfs/glusterd.log shows:
[2024-09-29 07:24:32.815997 +0000] E [MSGID: 106061] [glusterd-
snapshot-utils.c:986:gd_import_new_brick_snap_details] 0-management:
volume1.brick1.origin_path missing in payload 
[2024-09-29 07:24:32.816025 +0000] E [MSGID: 106061] [glusterd-
snapshot-utils.c:1004:gd_import_new_brick_snap_details] 0-management:
volume1.brick1.snap_type missing in payload 
[2024-09-29 07:24:32.924139 +0000] E [MSGID: 106061] [glusterd-
snapshot-utils.c:1093:gd_import_volume_snap_details] 0-management:
volume1.restored_from_snapname_id missing in payload for gv0 
[2024-09-29 07:24:32.924158 +0000] E [MSGID: 106061] [glusterd-
snapshot-utils.c:1104:gd_import_volume_snap_details] 0-management:
volume1.restored_from_snapname missing in payload for gv0 
[2024-09-29 07:24:32.924175 +0000] E [MSGID: 106061] [glusterd-
snapshot-utils.c:1115:gd_import_volume_snap_details] 0-management:
volume1.snap_plugin missing in payload for gv0 
[2024-09-29 07:24:32.924365 +0000] E [MSGID: 106061] [glusterd-
snapshot-utils.c:986:gd_import_new_brick_snap_details] 0-management:
volume1.brick1.origin_path missing in payload 
[2024-09-29 07:24:32.924385 +0000] E [MSGID: 106061] [glusterd-
snapshot-utils.c:1004:gd_import_new_brick_snap_details] 0-management:
volume1.brick1.snap_type missing in payload 
[2024-09-29 07:24:43.516131 +0000] E [MSGID: 106048] [glusterd-op-
sm.c:1814:glusterd_op_stage_status_volume] 0-management: Failed to get
volinfo [{Volume=gv0}] 
[2024-09-29 07:24:43.516209 +0000] E [MSGID: 106301] [glusterd-op-
sm.c:5870:glusterd_op_ac_stage_op] 0-management: Stage failed on
operation 'Volume Status', Status : -1 
[2024-09-29 07:24:43.524640 +0000] E [MSGID: 106048] [glusterd-op-
sm.c:1814:glusterd_op_stage_status_volume] 0-management: Failed to get
volinfo [{Volume=gv1}] 
[2024-09-29 07:24:43.524683 +0000] E [MSGID: 106301] [glusterd-op-
sm.c:5870:glusterd_op_ac_stage_op] 0-management: Stage failed on
operation 'Volume Status', Status : -1 

The cluster is running Ubuntu 20.04 9.6-ubuntu1~focal1
The new node is running Ubuntu 24.04 11.1-4ubuntu0.1

I know there is version mismatch, according to General Upgrade
procedure and v11 release notes I understood one can upgrade directly
from 9 -> 11, which I am trying to do by adding one new node after one
another and moving bricks over to new nodes step by step.
Also there is kind of deadlock, too, as launchpad Gluster Ubuntu repo
does not have an installation target for v9 on Ubuntu 24.04 and v10
does not run due to python errors when installed via launchpad repo.
Checked firewall && apparmor && switches && cabling.

Any ideas, comments towards the problem or approach?

Many thanks.
A.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20240929/04dfd8e7/attachment.html>