Bug 762906 (GLUSTER-1174) - Replicate spawns read loops even after destination fails
Summary: Replicate spawns read loops even after destination fails
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: GLUSTER-1174
Product: GlusterFS
Classification: Community
Component: replicate
Version: 3.1-alpha
Hardware: All
OS: Linux
low
high
Target Milestone: ---
Assignee: Pranith Kumar K
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-07-20 08:20 UTC by Shehjar Tikoo
Modified: 2011-02-18 08:47 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed:
Regression: RTP
Mount Type: nfs
Documentation: DNR
CRM:
Verified Versions:

Attachments (Terms of Use)
Add an attachment (proposed patch, testcase, etc.)

Description Shehjar Tikoo 2010-07-20 05:21:23 UTC 
Log available at: dev:/share/tickets/1174
Comment 1 Shehjar Tikoo 2010-07-20 08:20:19 UTC 
During self-heal, if the destination node goes down again, replicate continues to spawn read loops for reading from the source node. This causes delays in serving IO requests from clients.
Comment 2 Pranith Kumar K 2010-08-02 06:17:45 UTC 
on AWS instances, read loops are being spawned even without nodes going up and down.
Comment 3 Pavan Vilas Sondur 2010-08-10 06:07:57 UTC 
Seen performance degradation from rc9 to rc10 nfs-beta branch.
Comment 4 Raghavendra G 2010-09-20 09:42:56 UTC 
The bug does not occur on nfs-beta. However in order to verify whether this is fixed on 3.1, bug 762692, which is still reproducible on 3.1 has to be fixed. Hence leaving this bug open.
Comment 5 Anand Avati 2011-02-18 04:33:00 UTC 
PATCH: http://patches.gluster.com/patch/6147 in master (cluster/afr: stop spawning self-heal loops when the self-heal fails)
Note You need to log in before you can comment on or make changes to this bug.