Bug 761884 (GLUSTER-152)

Summary: libglusterfsclient: DHT workaround is a major performance bottleneck
Product: [Community] GlusterFS Reporter: Shehjar Tikoo <shehjart>
Component: libglusterfsclientAssignee: Shehjar Tikoo <shehjart>
Status: CLOSED CURRENTRELEASE QA Contact:
Severity: low Docs Contact:
Priority: low    
Version: mainlineCC: gluster-bugs
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: Type: ---
Regression: RTNR Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Shehjar Tikoo 2009-07-19 05:50:45 UTC
BTW, Avati also observed performance increase from 2-5 MiB/s with the problem, to around 40 MiB/s with a temporary fix he introduced on a customer setup.

Comment 1 Shehjar Tikoo 2009-07-19 05:54:49 UTC
Proposed patches:
release-2.0: http://patches.gluster.com/patch/784/
mainline: http://patches.gluster.com/patch/785/

Comment 2 Shehjar Tikoo 2009-07-19 08:49:23 UTC
As reported by Avati.

In libgf_client_path_lookup, the workaround for DHT requirement for a 
lookup on / needs to be done only once when the xlator graph is inited.

Doing it on every path's lookup results in a major performance penalty when using distribute subvolumes upwards of 16, because a lookup for the root needs to be sent on every subvolume.

Comment 3 Anand Avati 2009-07-20 18:29:29 UTC
PATCH: http://patches.gluster.com/patch/785 in master (libglusterfsclient: Move DHT root-lookup workaround to init)

Comment 4 Anand Avati 2009-07-20 18:29:40 UTC
PATCH: http://patches.gluster.com/patch/784 in release-2.0 (libglusterfsclient: Move DHT root-lookup workaround to init)