Bug 787722 - VDSM: deadlock on SpmStart due to changes in lock release
Summary: VDSM: deadlock on SpmStart due to changes in lock release
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: oVirt
Classification: Retired
Component: vdsm
Version: unspecified
Hardware: x86_64
OS: Linux
urgent
urgent
Target Milestone: ---
: 3.1
Assignee: Dan Kenigsberg
QA Contact:
URL:
Whiteboard: storage
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2012-02-06 15:46 UTC by Dafna Ron
Modified: 2016-02-10 16:35 UTC (History)
7 users (show)

Fixed In Version: v4.9.4-32-g5c2b4ae
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed:
oVirt Team: Storage


Attachments (Terms of Use)
logs (1.02 MB, application/x-gzip)
2012-02-06 15:49 UTC, Dafna Ron
no flags Details

Description Dafna Ron 2012-02-06 15:46:12 UTC
Description of problem:

we are getting a deadlock on SpmStart when SD is manually activated while SD is blocked from host (with iptables). 
because resource is not released host is stuck in contending and we are unable to squire SPM.  

Saggi checked and code change in commit 4d1dbcdd859be7a16ca0a276076b4a9f6ad5e7fa  "Fix lock release and reversed SPM logic"

self.lock should  be an RLOCK

Version-Release number of selected component (if applicable):

vdsm-4.9.3.2-0.fc16.x86_64

How reproducible:

100%

Steps to Reproduce:
1. create SD from 2 storage servers
2. block connectivity to 1 of the domains from host using iptables
3. after domain become inactive try to manually activate it
4. remove iptables block from host and try to activate domain again
  
Actual results:

deadlock in SpmStart, host will be stuck in contending and we cannot recover. 

Expected results:

we should be able to recover. 

Additional info: vdsm logs+ engine log

Comment 1 Saggi Mizrahi 2012-02-06 15:47:53 UTC
http://gerrit.ovirt.org/#change,1683

Comment 2 Dafna Ron 2012-02-06 15:49:44 UTC
Created attachment 559666 [details]
logs

Comment 3 Itamar Heim 2012-08-09 08:03:32 UTC
closing ON_QA bugs as oVirt 3.1 was released:
http://www.ovirt.org/get-ovirt/

Comment 4 Itamar Heim 2012-08-09 08:04:12 UTC
closing ON_QA bugs as oVirt 3.1 was released:
http://www.ovirt.org/get-ovirt/


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