Re: concurrent resource group problem.
- From: "unixsystems@xxxxxxxxx" <unixsystems@xxxxxxxxx>
- Date: 10 Jun 2006 08:12:22 -0700
Hi Garry,
It sounds like you have the second node volume group attributes set to
auto varyon. How did you create the VG? Did you create it as an
enhanced concurrent volume group, with autovaryon=false on node 1, then
import it onto node 2? If you did this then the attributes for the VG
would be in sync for both nodes.
Can please you do a 'lsvg volume_group' & 'lsattr -El volume_group' on
both nodes and post the output here. If it is true that node 2 is set
to auto varyon then it will be easy to fix - 'chvg -a n volume_group'.
Does the cluster successfully complete 'syncronisation and
verification'?
Garry191 wrote:
Hi folks, wonder if anyone can help, I'm not too experienced with HACMP
and have an urgent problem. Essentially we have a 2 node cluster
(AIX5.3L ML03). When the first node starts, it's concurrent volume
group is bought on-line correctly which contains a bunch of raw logical
volumes used by our app. When the second node is bought on-line, the
first node looses access to the volume group. The second node reports
the LV contents as JFS (not raw)
which is when you see the errors LVM_QUORUMNOQUORUM and
LVM_MISSPVADDED. If we varyoff and varyon the volume group on the first
node. everything seams fine. I'm sure it's a simple thing! Your help
would be much appreciated. I can give more details if you need dumps
etc.
thanks in advance.
.
- Follow-Ups:
- Re: concurrent resource group problem.
- From: Garry191
- Re: concurrent resource group problem.
- From: Garry191
- Re: concurrent resource group problem.
- References:
- concurrent resource group problem.
- From: Garry191
- concurrent resource group problem.
- Prev by Date: ooRexx 3.1 BETA1
- Next by Date: Re: concurrent resource group problem.
- Previous by thread: concurrent resource group problem.
- Next by thread: Re: concurrent resource group problem.
- Index(es):
Relevant Pages
|
