Bug 831630 - GlusterFS: Volumes sometimes fail to start
Summary: GlusterFS: Volumes sometimes fail to start
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: glusterfs
Version: unspecified
Hardware: Unspecified
OS: Unspecified
low
high
Target Milestone: ---
: ---
Assignee: Shireesh
QA Contact: Sudhir D
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2012-06-13 13:24 UTC by Daniel Paikov
Modified: 2013-07-04 07:57 UTC (History)
6 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2012-11-16 09:24:12 UTC
Embargoed:


Attachments (Terms of Use)
glusterfs log (9.87 KB, application/x-compressed-tar)
2012-06-13 13:24 UTC, Daniel Paikov
no flags Details
vdsm.log (329.07 KB, application/x-compressed-tar)
2012-06-13 13:25 UTC, Daniel Paikov
no flags Details

Description Daniel Paikov 2012-06-13 13:24:58 UTC
Created attachment 591499 [details]
glusterfs log

When starting several volumes one after the other, one of the volumes failed to start. The same volume could be successfully started later on.

From vdsm.log:
MainProcess|Thread-102376::DEBUG::2012-06-13 12:34:04,463::__init__::1164::Storage.Misc.excCmd::(_log) '/usr/sbin/gluster --mode=script volume start vol3-distrep' (cwd None)
MainProcess|Thread-102376::DEBUG::2012-06-13 12:34:04,575::__init__::1164::Storage.Misc.excCmd::(_log) FAILED: <err> = 'operation failed\n'; <rc> = 255
Thread-102376::DEBUG::2012-06-13 12:34:04,577::BindingXMLRPC::880::vds::(wrapper) return volumeStart with {'status': {'message': 'Volume start failed\nerror: operation failed\nreturn code: 255', 'code': 4125, 'rc': 255, 'err': ['operation failed'], 'out': []}}
MainProcess|Thread-102375::DEBUG::2012-06-13 12:34:04,948::__init__::1164::Storage.Misc.excCmd::(_log) SUCCESS: <err> = ''; <rc> = 0
Thread-102375::DEBUG::2012-06-13 12:34:04,949::BindingXMLRPC::880::vds::(wrapper) return volumeStart with {'status': {'message': 'Done', 'code': 0}}

From gluster log:
[2012-06-13 12:34:04.566422] I [glusterd-volume-ops.c:290:glusterd_handle_cli_start_volume] 0-glusterd: Received start vol reqfor volume vol3-distrep
[2012-06-13 12:34:04.566505] E [glusterd-utils.c:277:glusterd_lock] 0-glusterd: Unable to get lock for uuid: 876aef78-4863-470c-b9d1-ac912993c392, lock held by: 876aef78-4863-470c-b9d1-ac912993c392
[2012-06-13 12:34:04.566525] E [glusterd-handler.c:453:glusterd_op_txn_begin] 0-management: Unable to acquire local lock, ret: -1
[2012-06-13 12:34:04.944327] I [glusterd-rpc-ops.c:1316:glusterd3_1_commit_op_cbk] 0-glusterd: Received ACC from uuid: 9433c12e-2342-4663-9571-5587fa55300e
[2012-06-13 12:34:04.944397] I [glusterd-op-sm.c:2254:glusterd_op_modify_op_ctx] 0-management: op_ctx modification not required
[2012-06-13 12:34:04.946053] I [glusterd-rpc-ops.c:606:glusterd3_1_cluster_unlock_cbk] 0-glusterd: Received ACC from uuid: 9433c12e-2342-4663-9571-5587fa55300e
[2012-06-13 12:34:04.946088] I [glusterd-op-sm.c:2653:glusterd_op_txn_complete] 0-glusterd: Cleared local lock

Comment 1 Daniel Paikov 2012-06-13 13:25:20 UTC
Created attachment 591500 [details]
vdsm.log

Comment 3 Selvasundaram 2012-06-15 06:34:21 UTC
It seems to be GlusterFS bug.

Comment 4 Amar Tumballi 2012-09-18 05:04:12 UTC
handling multiple commands with gluster CLI is not supported at the moment, hence starting multiple volumes simultaneously failed.

Comment 5 Shireesh 2012-10-08 09:17:26 UTC
Daniel, Amar's statement conflicts with your statement that you were trying to start the volumes 'one after the other'. Can you please confirm whether you had tried to "select multiple volumes in the tab and start all of them at once" or "start one volume at a time, in a sequence"?

Comment 6 Daniel Paikov 2012-10-09 08:05:27 UTC
(In reply to comment #5)
> Daniel, Amar's statement conflicts with your statement that you were trying
> to start the volumes 'one after the other'. Can you please confirm whether
> you had tried to "select multiple volumes in the tab and start all of them
> at once" or "start one volume at a time, in a sequence"?

If I remember correctly, I started the volumes one volume at a time, via the GUI.

Comment 7 Shireesh 2012-10-09 08:26:48 UTC
Amar, in that case, it looks like a different issue, as the commands were fired one after the other?

Comment 8 Shireesh 2012-11-16 09:24:12 UTC
I'm closing this, assuming that it could be a corner case bug/limitation in glusterfs. RHS-C invokes the volume start verb on vdsm, which ultimately invokes the gluster cli command 'gluster volume start <volname>'. We can see clearly in the log that the command was invoked correctly, and the error came up from gluster cli. So it is not an issue with RHS-C. I feel gluster cli should provide more details on the error message than just 'operation failed'.


Note You need to log in before you can comment on or make changes to this bug.