Bug 1660701

Summary:	Use adaptive mutex in rpcsvc_program_register to improve performance
Product:	[Community] GlusterFS	Reporter:	Mohit Agrawal <moagrawa>
Component:	rpc	Assignee:	Mohit Agrawal <moagrawa>
Status:	CLOSED CURRENTRELEASE	QA Contact:
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	mainline	CC:	bugs, moagrawa, rgowdapp
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	glusterfs-6.0	Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2019-03-25 16:32:47 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:

Description Mohit Agrawal 2018-12-19 02:30:18 UTC

Description of problem:
Adaptive mutexes are used to protect critical/shared data items that are held for short periods.
It provides a balance between spin locks and traditional mutex.
We have observed after use adaptive mutex in rpcsvc_program_register got some improvement.

Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:

Comment 1 Worker Ant 2018-12-19 02:48:36 UTC

REVIEW: https://review.gluster.org/21883 (rpc: Use adaptive mutex in rpcsvc_program_register) posted (#1) for review on master by MOHIT AGRAWAL

Comment 2 Yaniv Kaul 2018-12-19 15:08:37 UTC

Is there upstream developer documentation on when they should be used and how?

Comment 3 Worker Ant 2018-12-20 16:45:06 UTC

REVIEW: https://review.gluster.org/21883 (rpc: Use adaptive mutex in rpcsvc_program_register) posted (#3) for review on master by Raghavendra G

Comment 4 Raghavendra G 2018-12-20 16:57:31 UTC

(In reply to Yaniv Kaul from comment #2)
> Is there upstream developer documentation on when they should be used and
> how?

Currently I am not aware of one. Even with [1] there was marginal regression due to [1]. Thinking it was due to the mutex controlling access to the request queue, I thought of using spinlocks as I was expecting critical section to be very small (just adding/removing an element to list). However, pthread conditional variables couldn't be used with spinlocks and hence as a second option we tried adaptive mutex. After running perf benchmarks (small file workload [3]) indeed showed adaptive mutex was helpful.

I referred to [4], which led to [5]

[1] https://review.gluster.org/r/I578c3fc67713f4234bd3abbec5d3fbba19059ea5
[2] https://review.gluster.org/17105
[3] https://github.com/distributed-system-analysis/smallfile
[4] https://stackoverflow.com/questions/19863734/what-is-pthread-mutex-adaptive-np
[5] http://lkml.iu.edu/hypermail/linux/kernel/0103.1/0030.html

Comment 5 Shyamsundar 2019-03-25 16:32:47 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-6.0, please open a new bug report.

glusterfs-6.0 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] https://lists.gluster.org/pipermail/announce/2019-March/000120.html
[2] https://www.gluster.org/pipermail/gluster-users/

Comment 6 Red Hat Bugzilla 2023-09-14 04:44:00 UTC

The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days