[Gluster-users] Cannot upgrade from 3.6.3 to 3.7.3

[Andreas Mather]

Hi All!

I wanted to do a rolling upgrade of gluster from 3.6.3 to 3.7.3, but after
the upgrade, the updated node won't connect.

The cluster has 4 nodes (vhost[1-4]) and 4 volumes (vol[1-4]) with 2
replicas each:
vol1: vhost1/brick1, vhost2/brick2
vol2: vhost2/brick1, vhost1/brick2
vol3: vhost3/brick1, vhost4/brick2
vol4: vhost4/brick1, vhost3/brick2

I'm trying to start the upgrade on vhost4. After restarting glusterd, peer
status shows all other peers as disconnected, the log has repeated entries
like this:

[2015-08-27 10:59:56.982254] E [MSGID: 106167]
[glusterd-handshake.c:2078:__glusterd_peer_dump_version_cbk] 0-management:
Error through RPC layer, retry again later
[2015-08-27 10:59:56.982335] E [rpc-clnt.c:362:saved_frames_unwind] (-->
/lib64/libglusterfs.so.0(_gf_log_callingfn+0x196)[0x7f1a7a9579e6] (-->
/lib64/libgfrpc.so.0(saved_frames_unwind+0x1de)[0x7f1a7a7229be] (-->
/lib64/libgfrpc.so.0(saved_frames_destroy+0xe)[0x7f1a7a722ace] (-->
/lib64/libgfrpc.so.0(rpc_clnt_connection_cleanup+0x9c)[0x7f1a7a72447c] (-->
/lib64/libgfrpc.so.0(rpc_clnt_notify+0x48)[0x7f1a7a724c38] )))))
0-management: forced unwinding frame type(GF-DUMP) op(NULL(2)) called at
2015-08-27 10:59:56.981550 (xid=0x2)
[2015-08-27 10:59:56.982346] W [rpc-clnt-ping.c:204:rpc_clnt_ping_cbk]
0-management: socket disconnected
[2015-08-27 10:59:56.982359] I [MSGID: 106004]
[glusterd-handler.c:5051:__glusterd_peer_rpc_notify] 0-management: Peer
<vhost3-int> (<72e2078d-1ed9-4cdd-aad2-c86e418746d1>), in state <Peer in
Cluster>, has disconnected from glusterd.
[2015-08-27 10:59:56.982491] W
[glusterd-locks.c:677:glusterd_mgmt_v3_unlock] (-->
/lib64/libglusterfs.so.0(_gf_log_callingfn+0x196)[0x7f1a7a9579e6] (-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(glusterd_mgmt_v3_unlock+0x541)[0x7f1a6f55ee91]
(-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(__glusterd_peer_rpc_notify+0x162)[0x7f1a6f4c6972]
(-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(glusterd_big_locked_notify+0x4c)[0x7f1a6f4bc90c]
(--> /lib64/libgfrpc.so.0(rpc_clnt_notify+0x90)[0x7f1a7a724c80] )))))
0-management: Lock for vol vol1 not held
[2015-08-27 10:59:56.982504] W [MSGID: 106118]
[glusterd-handler.c:5073:__glusterd_peer_rpc_notify] 0-management: Lock not
released for vol1
[2015-08-27 10:59:56.982608] W
[glusterd-locks.c:677:glusterd_mgmt_v3_unlock] (-->
/lib64/libglusterfs.so.0(_gf_log_callingfn+0x196)[0x7f1a7a9579e6] (-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(glusterd_mgmt_v3_unlock+0x541)[0x7f1a6f55ee91]
(-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(__glusterd_peer_rpc_notify+0x162)[0x7f1a6f4c6972]
(-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(glusterd_big_locked_notify+0x4c)[0x7f1a6f4bc90c]
(--> /lib64/libgfrpc.so.0(rpc_clnt_notify+0x90)[0x7f1a7a724c80] )))))
0-management: Lock for vol vol2 not held
[2015-08-27 10:59:56.982618] W [MSGID: 106118]
[glusterd-handler.c:5073:__glusterd_peer_rpc_notify] 0-management: Lock not
released for vol2
[2015-08-27 10:59:56.982728] W
[glusterd-locks.c:677:glusterd_mgmt_v3_unlock] (-->
/lib64/libglusterfs.so.0(_gf_log_callingfn+0x196)[0x7f1a7a9579e6] (-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(glusterd_mgmt_v3_unlock+0x541)[0x7f1a6f55ee91]
(-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(__glusterd_peer_rpc_notify+0x162)[0x7f1a6f4c6972]
(-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(glusterd_big_locked_notify+0x4c)[0x7f1a6f4bc90c]
(--> /lib64/libgfrpc.so.0(rpc_clnt_notify+0x90)[0x7f1a7a724c80] )))))
0-management: Lock for vol vol3 not held
[2015-08-27 10:59:56.982739] W [MSGID: 106118]
[glusterd-handler.c:5073:__glusterd_peer_rpc_notify] 0-management: Lock not
released for vol3
[2015-08-27 10:59:56.982844] W
[glusterd-locks.c:677:glusterd_mgmt_v3_unlock] (-->
/lib64/libglusterfs.so.0(_gf_log_callingfn+0x196)[0x7f1a7a9579e6] (-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(glusterd_mgmt_v3_unlock+0x541)[0x7f1a6f55ee91]
(-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(__glusterd_peer_rpc_notify+0x162)[0x7f1a6f4c6972]
(-->
/usr/lib64/glusterfs/3.7.3/xlator/mgmt/glusterd.so(glusterd_big_locked_notify+0x4c)[0x7f1a6f4bc90c]
(--> /lib64/libgfrpc.so.0(rpc_clnt_notify+0x90)[0x7f1a7a724c80] )))))
0-management: Lock for vol vol4 not held
[2015-08-27 10:59:56.982858] W [MSGID: 106118]
[glusterd-handler.c:5073:__glusterd_peer_rpc_notify] 0-management: Lock not
released for vol4
[2015-08-27 10:59:56.982881] W [socket.c:642:__socket_rwv] 0-management:
readv on 192.168.92.2:24007 failed (Connection reset by peer)
[2015-08-27 10:59:56.982974] E [rpc-clnt.c:362:saved_frames_unwind] (-->
/lib64/libglusterfs.so.0(_gf_log_callingfn+0x196)[0x7f1a7a9579e6] (-->
/lib64/libgfrpc.so.0(saved_frames_unwind+0x1de)[0x7f1a7a7229be] (-->
/lib64/libgfrpc.so.0(saved_frames_destroy+0xe)[0x7f1a7a722ace] (-->
/lib64/libgfrpc.so.0(rpc_clnt_connection_cleanup+0x9c)[0x7f1a7a72447c] (-->
/lib64/libgfrpc.so.0(rpc_clnt_notify+0x48)[0x7f1a7a724c38] )))))
0-management: forced unwinding frame type(GLUSTERD-DUMP) op(DUMP(1)) called
at 2015-08-27 10:59:56.981566 (xid=0x1)


Any ideas? Is it even possible to perform a rolling upgrade?

Thanks for any help!

Andreas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20150827/267ee981/attachment.html>