Operating System - HP-UX
1834052 Members
2270 Online
110063 Solutions
New Discussion

Re: lvextend & /etc/lvmtab in cluster

 
SOLVED
Go to solution
KVK
Valued Contributor

lvextend & /etc/lvmtab in cluster

I need the procedure of doing lvextend in Cluster .

Also i found in one node /etc/lvmtab is not having one VG entry but the same existing in another node /etc/lvmtab

The set right the above is it ok for me do a vgexport/vgimport ?

Awaiting from reply from experts soon
3 REPLIES 3
Warren_9
Honored Contributor

Re: lvextend & /etc/lvmtab in cluster

hi,

The lvextend in cluster env should be same as normal lvextend. i think it only need to perform the changes on one node.

for the extendfs, if the OnlineJFS is installed, you don't need to halt ths package to umount the FS. Otherwise, you need to stop the package.

GOOD LUCK!
Adisuria Wangsadinata_1
Honored Contributor
Solution

Re: lvextend & /etc/lvmtab in cluster

Hi,

Check the url below (docID : UXSGKBAN00000342) about 'Resizing a logical volume in a ServiceGuard package' :

http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000081631752

For your lvmtab issue, check the url below (docID : UMCSGKBRC00010012) about 'Repairing /etc/lvmtab on a ServiceGuard cluster node' :

http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000080090646

Hope this information can help you.

Cheers,
AW
now working, next not working ... that's unix
A. Clay Stephenson
Acclaimed Contributor

Re: lvextend & /etc/lvmtab in cluster

AS long as all you are doing is an lvextend then you can do this on just one node. Al.l the changes take place in the VGRA on the disks themselves so that as soon as the VG is activated on another node, this new node "knows" about them.

It is unclear about what you are saying about the lvmtab entries. The lvmtab entries on all the shared VG's do not have to be identical but they must be equivalent. For example, vg04 might have /dev/dsk/c1t5d0 and /dev/dsk/c1t4d0 on one node and vg04 might have /dev/dsk/c7t5d0 and /dev/dsk/c7t4d0 on another nodes. As long as these refer to the same physical devices, all is well.
If it ain't broke, I can fix that.