Bug 1558994 - 47% regression in mkdir from 3.3.1
Summary: 47% regression in mkdir from 3.3.1
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat
Component: arbiter
Version: rhgs-3.4
Hardware: Unspecified
OS: All
Target Milestone: ---
: RHGS 3.4.0
Assignee: Karthik U S
QA Contact: Karan Sandha
Depends On:
Blocks: 1503137
TreeView+ depends on / blocked
Reported: 2018-03-21 13:20 UTC by Karan Sandha
Modified: 2018-09-24 06:51 UTC (History)
8 users (show)

Fixed In Version: glusterfs-3.12.2-13
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Last Closed: 2018-09-04 06:44:20 UTC
Target Upstream Version:

Attachments (Terms of Use)

System ID Private Priority Status Summary Last Updated
Red Hat Product Errata RHSA-2018:2607 0 None None None 2018-09-04 06:45:35 UTC

Description Karan Sandha 2018-03-21 13:20:36 UTC
Description of problem:
Regression seen on smallfile mkdir from rhgs 3.3.1 to rhgs 3.4.0 on fuse mount. 

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

How reproducible:
profile infos are kept at sosreports are placed at sosreport and statedumps at mount -t cifs -oguest
//newtssuitcase.lab.eng.blr.redhat.com/newtssuitcase /BZ_sosreports/Karan/Porfile_info/

Steps to Reproduce:
1. Run small-file creates on 6 node setup using 6 client machines below command qeSyncDropCaches.sh; python /small-files/smallfile/smallfile_cli.py --operation mkdir --threads 8 --file-size 64 --files 5000 --top /gluster-mount  --host-set "`echo $CLIENT | tr ' ' ','`"
2. now compare the same with RHGS 3.3.1 i.e 

Actual results:
Regression seen between release  

Expected results:
No regression should be observed

Additional info:
not a brick mux setup.
The performance numbers weren't consistent i.e after every iterations the regression was more. 
regression seen on 2x2 volume also

Comment 3 Karan Sandha 2018-03-22 07:16:30 UTC
Profile infos at

mount -t cifs -oguest //newtssuitcase.lab.eng.blr.redhat.com/newtssuitcase


Comment 7 Karthik U S 2018-03-23 09:52:11 UTC

Please provide the strace output of the brick processes.

Comment 10 Raghavendra G 2018-04-16 06:00:38 UTC
I remember from an IRC conversation with Karan that this regression was not seen on plain distribute volumes. @Karan, can you please confirm that?

Comment 14 Sunil Kumar Acharya 2018-06-27 03:06:03 UTC
Upstream patch: https://review.gluster.org/#/c/20392/

Comment 19 errata-xmlrpc 2018-09-04 06:44:20 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.


Note You need to log in before you can comment on or make changes to this bug.