Bug 1547888

Summary:	[brick-mux] incorrect event-thread scaling in server_reconfigure()
Product:	[Community] GlusterFS	Reporter:	Milind Changire <mchangir>
Component:	core	Assignee:	Milind Changire <mchangir>
Status:	CLOSED CURRENTRELEASE	QA Contact:
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	mainline	CC:	amukherj, bugs, moagrawa, rgowdapp
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	glusterfs-v4.1.0	Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:
Clones:	1554255 1558959 (view as bug list)		Environment:
Last Closed:	2018-06-20 18:01:20 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	1554255, 1558959

Description Milind Changire 2018-02-22 08:36:30 UTC

Description of problem:
Whenever BRICK_ATTACH or BRICK_TERMINATE management operation is triggered on the brick, glusterfs_autoscale_threads() scales the event-threads correctly. However, the subsequent xlator_reconfigure()s that are invoked pass a default value of "event-threads" option and value to server_reconfigure() causing a rescaling of event-threads to 2. This is because the value of "event-threads" option is always 1, presuming that the "server.event-threads" wasn't set to a different value.

eg.
In case of brick-mux, event-threads are scaled by +1 for BRICK_ATTACH and scaled by -1 for BRICK_TERMINATE. So starting with 5 bricks, if 5 more bricks are added in a single operation, the brick receives 5 BRICK_ATTACH messages causing the event-threads to be incrementally scaled to 10. However, the xlator reconfigures that get invoked causes the system default value for "event-thread" option to be used causing the scaling of event-threads back to 2.

This may cause an unwanted fop throttling at the bricks.
Care needs to be taken so that the event thread count is not reset after rescaling at BRICK_ATTACH/BRICK_TERMINATE management op.

Comment 1 Worker Ant 2018-03-09 10:51:48 UTC

REVIEW: https://review.gluster.org/19689 (rpcsvc: correct event-thread scaling) posted (#1) for review on master by Milind Changire

Comment 2 Worker Ant 2018-03-12 08:49:13 UTC

COMMIT: https://review.gluster.org/19689 committed in master by "Raghavendra G" <rgowdapp> with a commit message- rpcsvc: correct event-thread scaling

Problem:
Auto thread count derived from the number of attachs and detachs
was reset to 1 when server_reconfigure() was called.

Solution:
Avoid auto-thread-count reset to 1.

Change-Id: Ic00e86adb81ba3c828e354a6ccb638209ae58b3e
BUG: 1547888
Signed-off-by: Milind Changire <mchangir>

Comment 3 Shyamsundar 2018-06-20 18:01:20 UTC

This bug is getting closed because a release has been made available that should address the reported issue. In case the problem is still not fixed with glusterfs-v4.1.0, please open a new bug report.

glusterfs-v4.1.0 has been announced on the Gluster mailinglists [1], packages for several distributions should become available in the near future. Keep an eye on the Gluster Users mailinglist [2] and the update infrastructure for your distribution.

[1] http://lists.gluster.org/pipermail/announce/2018-June/000102.html
[2] https://www.gluster.org/pipermail/gluster-users/