Bug 764782 (GLUSTER-3050) - 'replace-brick' hangs on vm's
Summary: 'replace-brick' hangs on vm's
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: GLUSTER-3050
Product: GlusterFS
Classification: Community
Component: unclassified
Version: 3.2.1
Hardware: x86_64
OS: Linux
medium
medium
Target Milestone: ---
Assignee: krishnan parthasarathi
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2011-06-17 11:28 UTC by M S Vishwanath Bhat
Modified: 2016-06-01 01:57 UTC (History)
4 users (show)

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


Attachments (Terms of Use)

Description M S Vishwanath Bhat 2011-06-17 11:28:05 UTC
I tried a basic replace-brick with the bare minimum data on the source-brick. But the replace-brick is hung and 'replace-brick status' returns "Migration is in progress" every time. 

#gluster volume replace-brick hosdu 192.168.1.222:/tmp/brick2/ 192.168.1.222:/tmp/brick4/ start
replace-brick started successfully
#gluster volume replace-brick hosdu 192.168.1.222:/tmp/brick2/ 192.168.1.222:/tmp/brick4/ status
Number of files migrated = 0       Current file=  
sleep 20;
#gluster volume replace-brick hosdu 192.168.1.222:/tmp/brick2/ 192.168.1.222:/tmp/brick4/ status
Number of files migrated = 0       Current file=  

This issue is happening only for vm's with only single processor. With my laptop with 4 processors this issue is not happening.

Comment 1 Anand Avati 2011-06-20 04:21:08 UTC
PATCH: http://patches.gluster.com/patch/7555 in master (pump: mark pending before notify to children to avoid race in single CPU.)

Comment 2 krishnan parthasarathi 2011-06-20 09:55:55 UTC
Steps to verify the fix:
- Try performing replace-brick operation in a single CPU machine/vm.

Comment 3 Anand Avati 2011-06-22 12:41:25 UTC
PATCH: http://patches.gluster.com/patch/7572 in release-3.1 (pump: mark pending before notify to children to avoid race in single CPU.)

Comment 4 Anand Avati 2011-06-22 12:42:07 UTC
PATCH: http://patches.gluster.com/patch/7571 in release-3.2 (pump: mark pending before notify to children to avoid race in single CPU.)

Comment 5 Saurabh 2011-07-01 02:46:13 UTC
what is the target milestone

Comment 6 M S Vishwanath Bhat 2011-07-04 06:40:41 UTC
This is fixed. I tried with 3.2.2qa1 and it's working fine.

Comment 7 M S Vishwanath Bhat 2011-07-29 08:50:25 UTC
Ran the verify pump script on a single cpu vm which had 3.1.6qa2. Issuse is fixed now in release-3.1 branch also.


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