Note: This is a beta release of Red Hat Bugzilla 5.0. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Also email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback here.
Bug 1513721 - Evaluate/update clustered mirrored volume procedure in LVM manual
Summary: Evaluate/update clustered mirrored volume procedure in LVM manual
Status: NEW
Alias: None
Product: Red Hat Enterprise Linux 7
Classification: Red Hat
Component: doc-Logical_Volume_Manager
Version: 7.4
Hardware: Unspecified
OS: Unspecified
Target Milestone: rc
: ---
Assignee: Steven J. Levine
QA Contact: Corey Marthaler
Depends On:
TreeView+ depends on / blocked
Reported: 2017-11-15 20:45 UTC by Steven J. Levine
Modified: 2019-03-06 00:43 UTC (History)
6 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Last Closed:
Target Upstream Version:

Attachments (Terms of Use)

Description Steven J. Levine 2017-11-15 20:45:08 UTC
The clustered mirror example in the LVM manual is a RHEL 7 update (using Pacemaker) of what had been a clustered mirror example in RHEL 6 which was meant to show both mirrored configuration an cluster configuration.  But we now have other (more thorough) procedure examples of configuring LVM volumes in a cluster and we also consider RAID mirrors to be preferable (although they are not supported in clusters).  

This BZ is a memo to look again at that procedure to determine what it is we want to desmonstrate now and whether this is the best way to do it.

Some info taken from BZ#1507973, from John Ruemker:


> 3.
> The LVM manual includes a procedure for configured a mirrored volume in
> a cluster.  This procedure was originally (pre-Pacemaker) meant to show
> an example of LVM in a cluster and also to show an example of
> configuring a mirrored volume.  I had to update the procedure for RHEL 7
> to include the Pacemaker info.  But the mirror example here is an
> old-style mirror and if the point is to show a mirror example it really
> should be a RAID volume.  So I will replace that perhaps with a
> different example, using a RAID volume, but I'm not sure if we need to
> call out anything special about a RAID volume in a cluster -- are there
> resource agent parameters you need to use, for example?  It seems a
> different issue than the rest of this BZ so for now I have cloned that
> piece out into it's own BZ and I will pursue that separately.

RAID mirrors use the same LVM agent and require no special configuration in
the cluster - things should just work if you're using HA-LVM with tagging.
 (Although there is one resource attribute that you can modify that has
relevance with mirrors - partial_activation; setting to true allows for
activating a volume group with some PVs missing, which is a reasonable
expectation if you're using a mirror volume to handle partial failures;
 keep in mind this setting is applicable for LVM in general, so both types
of mirrors and non-mirrors also).

With CLVMD, you can only use old-style mirrors.

> The problematic section is section 5.5 here:
> enterprise_linux/7/html/logical_volume_manager_
> administration/mirvol_create_ex
> Sorry for the length/complexity of this note, but I'm really trying to
> simplify things.

"# pcs resource update clvmd with_cmirrord=true"

We have or had a bug open dealing with the fact that the clvm agent doesn't
handle this update gracefully, but rather restarts clvmd to apply it.  So,
you need a downtime to do this.  Might be worth mentioning - although
probably also good to dig up that bug and see what we did with it.  We
might have maintained that expectation, or fixed it to go without the
restart - I can't recall.


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