Back to bug 1017007

Who When What Removed Added
Shruti Sampat 2013-10-09 07:12:43 UTC Severity unspecified high
Sudhir D 2013-10-09 09:23:26 UTC Keywords ZStream
Prasanth 2013-10-09 12:15:27 UTC Severity high urgent
Kaushal 2013-10-10 12:10:01 UTC Status NEW MODIFIED
CC kaushal
Matt Mahoney 2013-10-10 14:10:25 UTC Priority unspecified urgent
Vivek Agarwal 2013-10-11 04:00:33 UTC CC vagarwal
Krutika Dhananjay 2013-10-17 04:53:22 UTC Status MODIFIED ON_QA
Fixed In Version glusterfs-3.4.0.35rhs
Shruti Sampat 2013-10-18 09:30:30 UTC Status ON_QA VERIFIED
Krutika Dhananjay 2013-10-28 06:09:32 UTC Doc Text Cause:
During UNLOCK phase of a glusterd volume transaction, if the glusterd lock is found not to be taken, the control goes to 'out' where 'conf' is dereferenced even before it is initialised to a valid address.

Consequence:
The originator glusterd crashes when any two volume transactions are executed concurrently.

Fix:
Initialise 'conf' to a valid address before doing anything else as part of UNLOCK phase.

Result:
Any two concurrent volume ops executed concurrently will no longer cause glusterd to crash.
Anjana Suparna Sriram 2013-11-08 10:01:38 UTC CC asriram
Doc Text Cause:
During UNLOCK phase of a glusterd volume transaction, if the glusterd lock is found not to be taken, the control goes to 'out' where 'conf' is dereferenced even before it is initialised to a valid address.

Consequence:
The originator glusterd crashes when any two volume transactions are executed concurrently.

Fix:
Initialise 'conf' to a valid address before doing anything else as part of UNLOCK phase.

Result:
Any two concurrent volume ops executed concurrently will no longer cause glusterd to crash.
Previously, due to locking issues in glusterd management deamon crashed when two volume operations were executed simultaneously on the same node. Now, with this update this issue has been fixed.
Anjana Suparna Sriram 2013-11-11 10:36:29 UTC Doc Text Previously, due to locking issues in glusterd management deamon crashed when two volume operations were executed simultaneously on the same node. Now, with this update this issue has been fixed. Previously, due to locking issues in glusterd management deamon, glusterd crashed when two volume operations were executed simultaneously on the same node. Now, with this update this issue has been fixed.
errata-xmlrpc 2013-11-27 15:41:45 UTC Status VERIFIED CLOSED
Resolution --- ERRATA
Last Closed 2013-11-27 10:41:45 UTC

Back to bug 1017007