Back to bug 1327751

Who When What Removed Added
Red Hat Bugzilla Rules Engine 2016-04-15 20:20:31 UTC Keywords ZStream
Dustin Black 2016-04-15 20:26:14 UTC Keywords ZStream
Priority unspecified high
Kaushal 2016-04-18 11:31:42 UTC CC kaushal
Riyas Abdulrasak 2016-04-26 11:23:50 UTC CC rnalakka
Riyas Abdulrasak 2016-04-26 13:09:00 UTC Flags needinfo?(amukherj)
Kaushal 2016-04-28 08:27:27 UTC Depends On 1331289
Atin Mukherjee 2016-04-29 11:26:50 UTC Flags needinfo?(amukherj)
Atin Mukherjee 2016-04-29 11:29:13 UTC Status NEW POST
Assignee amukherj kaushal
Atin Mukherjee 2016-05-03 13:54:57 UTC Keywords ZStream
Rahul Hinduja 2016-05-04 12:03:25 UTC CC rhinduja
Blocks 1311817
Alok 2016-05-04 12:04:52 UTC CC asrivast
Rejy M Cyriac 2016-05-04 18:46:19 UTC CC rcyriac
Red Hat Bugzilla Rules Engine 2016-05-05 00:16:12 UTC Target Release --- RHGS 3.1.3
Atin Mukherjee 2016-05-05 05:53:34 UTC Status POST MODIFIED
Rahul Hinduja 2016-05-09 14:57:44 UTC QA Contact storage-qa-internal bsrirama
errata-xmlrpc 2016-05-10 11:26:20 UTC Status MODIFIED ON_QA
Milind Changire 2016-05-10 12:04:24 UTC Fixed In Version glusterfs-3.7.9-4
Rejy M Cyriac 2016-05-10 16:20:48 UTC Flags needinfo?(asrivast)
Alok 2016-05-11 05:56:01 UTC Flags needinfo?(asrivast)
Kaushal 2016-05-11 06:38:41 UTC Doc Text Cause:
Encrypted connection in GlusterFS are handled using a separate thread per connection. These threads were not cleaned up after disconnection.

Consequence:
This lead to a gradual increase in memory consumption of GlusterFS processes. This was particularly visible with GlusterD, which gets a lot of short lived connections.

Fix:
The dead threads are now cleaned up, once every minute.

Result:
The constant increase in memory consumption is avoided.
Rejy M Cyriac 2016-05-11 07:57:19 UTC CC sankarshan
Flags needinfo?(sankarshan)
sankarshan 2016-05-11 11:47:15 UTC Flags needinfo?(sankarshan)
Byreddy 2016-05-17 06:23:56 UTC Status ON_QA VERIFIED
Bipin Kunal 2016-05-30 08:42:55 UTC CC bkunal
Bipin Kunal 2016-05-30 08:45:28 UTC Status VERIFIED ON_QA
Rahul Hinduja 2016-05-30 09:52:32 UTC Status ON_QA VERIFIED
Bipin Kunal 2016-05-30 10:16:21 UTC CC annair
Flags needinfo?(annair)
Laura Bailey 2016-06-03 05:53:55 UTC Doc Text Cause:
Encrypted connection in GlusterFS are handled using a separate thread per connection. These threads were not cleaned up after disconnection.

Consequence:
This lead to a gradual increase in memory consumption of GlusterFS processes. This was particularly visible with GlusterD, which gets a lot of short lived connections.

Fix:
The dead threads are now cleaned up, once every minute.

Result:
The constant increase in memory consumption is avoided.
When encrypted connections are enabled, a separate thread is used for each connection. These threads were not being cleaned up correctly after the connections ended, which led to a gradual increase in GlusterFS memory consumption over time, especially in the glusterd process. These disconnected threads are now cleaned up once a minute, avoiding the problem with memory usage.
Flags needinfo?(kaushal)
Kaushal 2016-06-06 03:57:34 UTC Flags needinfo?(annair) needinfo?(kaushal)
errata-xmlrpc 2016-06-23 00:53:01 UTC Status VERIFIED RELEASE_PENDING
errata-xmlrpc 2016-06-23 05:18:02 UTC Status RELEASE_PENDING CLOSED
Resolution --- ERRATA
Last Closed 2016-06-23 01:18:02 UTC
Rejy M Cyriac 2016-09-17 16:34:16 UTC Sub Component glusterd
Component glusterd glusterd-transition
Rejy M Cyriac 2016-09-17 16:47:24 UTC CC rhs-bugs, storage-qa-internal, vbellur
Component glusterd-transition glusterd

Back to bug 1327751