cluster help

[leeji]

Hi.all
I found following message in syslog.log
WARNING: Cluster lock on disk /dev/dsk/c1t5d0 is missing!
Until it is fixed, a single failure could
cause all nodes in the cluster to crash
and
when I use cmapply command following appeared
Error: Volume group /dev/vg00 is configured differently on node dnappc2 than on node dnappc1
Error: Volume group /dev/vg00 on node dnappc1 does not appear to have a physical volume

corresponding to /dev/dsk/c0t5d0 on node dn.
Error: Volume group /dev/vg00 on node dnappc1 does not appear to have a physical volume

corresponding to /dev/dsk/c0t8d0 on node dn.
Do you ahve any idea?
model :D390 -2-
array:autoraid -1-
with inline scsi cable -4-
external scsi card -4-

[S.K. Chan]

Take a look at this document, it may be applicable to your problem.
DocID=KBRC00001982

http://support2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=500000000092231

Hope it helps..

[Frederic Sevestre]

Hi,

It seams that you tried to share the vg00. It is not possible.
Check the cluster ascii file to see if there is the vg00 (VOLUME_GROUP /dev/vg00).
In that case try cmapplyconf after deleting this line.

Fr??d??ric

[leeji]

No , thereis no VG00 in ascii file
Um, So I try to recompile cluster file using cmapply .
But same error appears!!
Does it need to reinstall OS?
Is there any solution or idea?
very interesting thing!!!

[Sanjay_6]

Hi,

What is the output of the vgdisplay for vg00 on server dnappc1.

vgdisplay -v /dev/vg00

Can you post the output of this vgdisplay for us to have a look. Also post the output (/tmp/cmclconf_chk) of this command.

cmcheckconf -v -C /etc/cmcluster/cmclconf.ascii >/tmp/cmchkconf.txt

Here cmclconf.ascii is your cluster conf file. If possible post the outpur of the cmquerycl command too,

cmquerycl -v -n node1 -n node2 >/tmp/cmqerycl.txt

Hope this helps.

regds

[Stephen Doud]

Hello Leeji,
The clusterlock disk issue can be resolved using this document: UXSGKBAN00000022
TITLE: "Cluster Lock On Disk {dsk} Is Missing"


I have seen the problem with vg00 when a boot disk of one node in the cluster was created from a "dd" of vg00 from another node in the cluster. The problem is that the VGID (volume group ID) is the SAME on both server's vg00 disks. When a modification is made to vg00 on one node, but not the other, and subsequently, a cmquerycl, cmcheckconf or cmapplyconf are performed, SG inspects all VG's in each /etc/lvmtab to see if they have the same VGID between nodes. If so, SG believes them to be one VG common to both nodes. After this, if it notices that the LVM structures of vg00 don't match between the nodes, though the VGID is the same - hence the errors you see.

If you used the 'dd' command to create the second server's boot disk, the ONLY resolution is to reinstall that disk. I have seen some customers do this quickly by using a make_recovery tape created from one server, and restore it to the other server... then customize it's hostname and IP etc.

This document discusses the issue:
UXSGKBRC00007573
TITLE: "Cloned VG00s and ServiceGuard configuration failures"

-s.

[Giri Sekar.]

Hi:

Try this..


1.halt the cluster

2. vgchange -a y /dev/vglock (activate the cluster lock on the node)

3. cmapplyconf -C /etc/cmcluster/cmclconf

-P /etc/cmcluster/pkgdir/pkgconf


4. vgchange -a n /dev/vglock(deactivate)

5. start the cluster.