Bug 1286026

Summary: partial_activation check fails in pacemaker
Product: Red Hat Enterprise Linux 6 Reporter: Josef Zimek <pzimek>
Component: resource-agentsAssignee: Oyvind Albrigtsen <oalbrigt>
Status: CLOSED CURRENTRELEASE QA Contact: cluster-qe <cluster-qe>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 6.7CC: agk, cfeist, cluster-maint, fdinitto, jruemker, pzimek
Target Milestone: rc   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-08-02 17:36:26 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1269194    

Description Josef Zimek 2015-11-27 09:38:59 UTC
Description of problem:
=======================
Customer is using a HA-LVM mirror logical volumes as resource for Pacemaker cluster.

When using the option 'partial_activation=true', like in the following resource,

# pcs resource show vgexample
 Resource: vgexample (class=ocf provider=heartbeat type=LVM)
  Attributes: volgrpname=vg_example exclusive=true partial_activation=true
 ...

Customer expects that in case of failure of a logical volume segment (missing PV), the cluster can still activate the VG and start the volume (in the LVM script, 'partial_activation=true' converts to 'vgchange -ay ... <vg> --partial')


But when running:

# pcs resource move res_grp    # res_grp is resource group with the vgexample resource

it fails.


It's because in the LVM ocf script, before doing the vgchange, the script
runs 'vgck <vg>' (in the LVM_validate_all() function) which fails because of
the missing PV.

When commenting out the vgck part as a test and everything worked as expected.





Version-Release number of selected component (if applicable):
=============================================================
RHEL 6.7

# rpm -qf /usr/lib/ocf/resource.d/heartbeat/LVM
resource-agents-3.9.5-24.el6.x86_64



How reproducible:
=================
always


Steps to Reproduce:
===================
1. move resource group which consist of LVM resource with partial_activation=true


Actual results:
===============
Resource group fails to move


Expected results:
=================
The LVM_validate_all() should check the value of OCF_RESKEY_partial_activation
before trying vgck. And the resource group moves to other node


Additional info:

Comment 2 Oyvind Albrigtsen 2015-11-27 09:54:33 UTC
Can he test the patch I've attached to the following bz?
https://bugzilla.redhat.com/show_bug.cgi?id=1266173

Comment 3 Oyvind Albrigtsen 2015-12-16 14:38:33 UTC
Josef: Can you ask the customer if he can test the patch mentioned in comment #2?

Comment 4 Josef Zimek 2015-12-16 15:03:46 UTC
Oyvind: Yes, I will let you know the result once I got feedback from customer. Thanks

Comment 7 John Ruemker 2016-08-02 17:36:26 UTC
Closing this out.  The issue described is resolved via the errata that was released in 6.8 via Bugzilla #1266173.  The customer case eventually closed without confirmation of the fix, but we expect that it should have been resolved.