Bug 1304287

Summary: Make the WRITE_IS_APPEND feature in AFR dynamically configurable for performance necessity
Product: [Community] GlusterFS Reporter: Ping.Lee <li.ping288>
Component: replicateAssignee: Ping.Lee <li.ping288>
Status: CLOSED DEFERRED QA Contact:
Severity: low Docs Contact:
Priority: medium    
Version: mainlineCC: bugs, li.ping288, ravishankar, smohan
Target Milestone: ---Keywords: Triaged
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2018-11-20 06:06:31 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 Ping.Lee 2016-02-03 09:15:24 UTC
Description of problem:

    Problem:
    AS discussed at
    http://www.gluster.org/pipermail/gluster-devel/2016-January/047915.html
    
    The WRITE_IS_APPEND in afr_writev leads to locking for each file writing operation
    in posix_writev in the server end.  This lock then impacts on the concurrent ability
    of posix_writev to do final file writing and results in AFR writing performance drop.

    Fix:
    Make the WRITE_IS_APPEND feature in AFR dynamically configurable.
    Enable or disable it under the necessity of performance.

Comment 1 Vijay Bellur 2016-02-05 02:34:02 UTC
REVIEW: http://review.gluster.org/13359 (cluster/afr: Make the WRITE_IS_APPEND feature dynamically              configurable for performance necessity) posted (#2) for review on master by Ping.Lee (li.ping288.cn)

Comment 2 Vijay Bellur 2016-02-05 02:43:41 UTC
REVIEW: http://review.gluster.org/13359 (cluster/afr: Make the WRITE_IS_APPEND feature dynamically configurable for performance necessity) posted (#3) for review on master by Ping.Lee (li.ping288.cn)

Comment 3 Vijay Bellur 2016-02-05 07:52:47 UTC
REVIEW: http://review.gluster.org/13364 (cluster/afr: Make the WRITE_IS_APPEND feature dynamically configurable for performance necessity) posted (#1) for review on master by Ping.Lee (li.ping288.cn)

Comment 4 Vijay Bellur 2016-02-05 08:08:31 UTC
REVIEW: http://review.gluster.org/13359 (cluster/afr: Make the WRITE_IS_APPEND feature dynamically configurable for performance necessity) posted (#4) for review on master by Ping.Lee (li.ping288.cn)

Comment 5 Mike McCune 2016-03-28 22:55:34 UTC
This bug was accidentally moved from POST to MODIFIED via an error in automation, please see mmccune with any questions

Comment 6 Ravishankar N 2018-11-20 06:06:31 UTC
Closing the bug as there is no activity and the patch has been abandoned. Please re-open it and refresh the patch if the fix is still needed.