Back to bug 1306194

Who When What Removed Added
Red Hat Bugzilla Rules Engine 2016-02-10 09:45:45 UTC Keywords ZStream
Nag Pavan Chilakam 2016-02-10 09:46:20 UTC Priority unspecified urgent
Mohammed Rafi KC 2016-02-10 16:13:31 UTC CC rkavunga
Nagaprasad Sathyanarayana 2016-02-11 11:47:47 UTC CC nsathyan
Assignee rhs-bugs rkavunga
Soumya Koduri 2016-02-11 14:00:05 UTC CC skoduri
Raghavendra G 2016-02-12 06:43:36 UTC CC rgowdapp
Mohammed Rafi KC 2016-02-12 11:21:35 UTC Blocks 1306930
Rejy M Cyriac 2016-02-12 12:58:57 UTC CC rcyriac
Rejy M Cyriac 2016-02-12 13:02:48 UTC Blocks 1268895
Laura Bailey 2016-02-15 02:45:57 UTC Flags needinfo?(rkavunga)
Mohammed Rafi KC 2016-02-15 06:15:28 UTC Flags needinfo?(rkavunga)
Laura Bailey 2016-02-16 01:07:44 UTC Blocks 1268895
Bhaskarakiran 2016-02-16 05:54:39 UTC CC byarlaga
Blocks 1299184
John Skeoch 2016-02-18 00:23:42 UTC CC nsathyan smohan
Joseph Elwin Fernandes 2016-02-18 09:13:16 UTC Whiteboard tier-fuse-nfs-samba
Mohammed Rafi KC 2016-02-23 07:29:25 UTC Blocks 1311002
Joseph Elwin Fernandes 2016-03-01 01:54:49 UTC Blocks 1299184
Joseph Elwin Fernandes 2016-04-15 12:59:10 UTC CC josferna
Alok 2016-04-18 09:38:31 UTC CC asrivast
Rahul Hinduja 2016-04-18 11:26:01 UTC CC rhinduja
Blocks 1299184
Red Hat Bugzilla Rules Engine 2016-04-18 11:51:35 UTC Target Release --- RHGS 3.1.3
Rahul Hinduja 2016-04-27 09:10:29 UTC QA Contact nchilaka kramdoss
Mohammed Rafi KC 2016-04-28 09:51:16 UTC Status NEW POST
Mohammed Rafi KC 2016-05-06 05:28:25 UTC Blocks 1333645
Nithya Balachandran 2016-05-06 15:39:01 UTC Status POST MODIFIED
CC nbalacha
errata-xmlrpc 2016-05-10 11:25:47 UTC Status MODIFIED ON_QA
Milind Changire 2016-05-10 12:04:37 UTC Fixed In Version glusterfs-3.7.9-4
Nithya Balachandran 2016-05-20 15:06:37 UTC Doc Text Cause:
DHT sends an unlock request to the wrong brick after an attach tier causing the lock to be held forever. All operations synchronised on that lock are blocked causing the mount point to hang.


Consequence:
All operations synchronised on that lock are blocked causing the mount point to hang.

Fix:
DHT now sends the unlock to the correct brick.

Result:
I/O proceeds after an attach tier without hanging.
krishnaram Karthick 2016-05-30 14:42:58 UTC Status ON_QA VERIFIED
Laura Bailey 2016-06-03 04:00:04 UTC Doc Text Cause:
DHT sends an unlock request to the wrong brick after an attach tier causing the lock to be held forever. All operations synchronised on that lock are blocked causing the mount point to hang.


Consequence:
All operations synchronised on that lock are blocked causing the mount point to hang.

Fix:
DHT now sends the unlock to the correct brick.

Result:
I/O proceeds after an attach tier without hanging.
When attach tier occurred in parallel with I/O, it was possible for the cached subvolume to change. This meant that if an I/O lock had been set on the cached volume just before the cached subvolume changed, unlock operations were sent to the wrong brick, and the lock on the original brick was never released. The location of the last lock is now recorded so that this issue no longer occurs even if the cached subvolume does change during these simultaneous operations.
Flags needinfo?(rkavunga)
Mohammed Rafi KC 2016-06-03 06:12:00 UTC Flags needinfo?(rkavunga)
Mohammed Rafi KC 2016-06-17 07:25:35 UTC Blocks 1347524
John Skeoch 2016-06-20 00:02:25 UTC CC josferna sankarshan
errata-xmlrpc 2016-06-23 00:48:13 UTC Status VERIFIED RELEASE_PENDING
errata-xmlrpc 2016-06-23 05:07:24 UTC Status RELEASE_PENDING CLOSED
Resolution --- ERRATA
Last Closed 2016-06-23 01:07:24 UTC
Rejy M Cyriac 2016-09-17 15:44:27 UTC Sub Component tiering
CC rhs-bugs, storage-qa-internal
Component glusterfs tier

Back to bug 1306194