Bug 763134 (GLUSTER-1402)

Summary: bonnie++ fails over cascading replicated backends
Product: [Community] GlusterFS Reporter: Anush Shetty <anush>
Component: replicateAssignee: Pranith Kumar K <pkarampu>
Status: CLOSED WONTFIX QA Contact:
Severity: medium Docs Contact:
Priority: low    
Version: 3.1-alphaCC: amarts, gluster-bugs, vijay, vijaykumar
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: fuse
Documentation: DNR CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
Client log file none

Description Anush Shetty 2010-08-19 12:47:00 UTC 
When I tried running bonnie++ over 3 subvolume replicate setup, and also bring the servers up and down continuously triggering self-heal, it exits out.

/mnt/gluster$ bonnie++ -u anush -d bonnie/
Using uid:1000, gid:1000.
Writing with putc()...Can't putc() - disk full?
Comment 1 Amar Tumballi 2011-03-22 10:04:38 UTC 
Need to reproduce with mainline first before saying there are issues..
Comment 2 Amar Tumballi 2011-04-25 09:33:25 UTC 
Please update the status of this bug as its been more than 6months since its filed (bug id < 2000)

Please resolve it with proper resolution if its not valid anymore. If its still valid and not critical, move it to 'enhancement' severity.
Comment 3 Vijaykumar 2011-08-17 10:14:36 UTC 
I did the similar test ,that i tried running bonnie++ on the fuse mount of 3 replica volume, also brought 2 servers up and down and kept the third server as it is, but bonnie++ did fail. Bonnie++ will fail only when you bring up the two servers and you immediately bring down the third before the first two get synced with the third, which is actually the case of split-brain condition.
Comment 4 Pranith Kumar K 2011-08-17 10:15:43 UTC 
(In reply to comment #3)
> I did the similar test ,that i tried running bonnie++ on the fuse mount of 3
> replica volume, also brought 2 servers up and down and kept the third server as
> it is, but bonnie++ did fail. Bonnie++ will fail only when you bring up the two
> servers and you immediately bring down the third before the first two get
> synced with the third, which is actually the case of split-brain condition.

This is the design limitation of afr. So wont be fixing it.