Bug 1286292

Summary:	rgmanager - pacemaker LVM RA disparity in an ability to further limit the scope to specific LV on top of given VG
Product:	Red Hat Enterprise Linux 7	Reporter:	Jan Pokorný [poki] <jpokorny>
Component:	resource-agents	Assignee:	Oyvind Albrigtsen <oalbrigt>
Status:	CLOSED NOTABUG	QA Contact:	cluster-qe <cluster-qe>
Severity:	unspecified	Docs Contact:
Priority:	low
Version:	7.2	CC:	agk, cluster-maint, fdinitto, oalbrigt
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:	2016-01-21 11:58:51 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:

Description Jan Pokorný [poki] 2015-11-27 20:19:06 UTC

Currently, clufter (the stacks migration tool) cannot serve users that
had lvm resource defined specifically for selected LV pertaining given VG.

It can also be viewed as sort of a regression: something that was possible
in the past (with lvm.sh rgmanager agent) cannot be done anymore (with
LVM heartbeat agent).

(The same can be extrapolated to the whole set of RAs and their disparity,
i.e., some agents may be missed for those who migrate, but combo of that
agent with specified LV seems quite popular).

Comment 2 Jan Pokorný [poki] 2015-11-27 20:23:13 UTC

This possibly applies to new RHEL 6 minor releases as well.

Comment 4 Jan Pokorný [poki] 2015-11-30 15:15:23 UTC

From clufter perspective, we may may actually just raise the warning if
there is no more LVs for given VG referred to from another agent instance,
and have more accurate failure message otherwise.

Comment 5 Jan Pokorný [poki] 2015-12-01 18:45:48 UTC

Re clufter:
I am now of the opinion that failing the conversion whenever lvm/@lv_name
is observed is a sanest way to do as one cannot make any safe conclusions
just by the content of cluster.conf.

Comment 7 Oyvind Albrigtsen 2015-12-23 14:14:58 UTC

The rgmanager resource agent only support 1 LV pr VG, so this is not an issue:
Dec 23 15:08:59 rgmanager [lvm] HA LVM requires Only one logical volume per volume group.
Dec 23 15:08:59 rgmanager [lvm] There are currently 2 logical volumes in test_vg
Dec 23 15:08:59 rgmanager [lvm] Failing HA LVM start of test_vg/test_lv

Comment 8 Jan Pokorný [poki] 2016-10-20 17:51:01 UTC

Returned to this and have to wonder why description of lv_name states:

> Name of the logical volume being managed.  This    
> parameter is optional if there are more than one   
> logical volumes in the volume group to be managed. 

then?

Any idea why there's such a contradiction or if I am missing something,
Oyvind?

If the description is indeed misleading, we would be better off fixing
it so that I am (and in turn, clufter users) not the only victim.