drd clone and SAN replication on a second datacenter

[lucas_2]

Hello,

I have 2 identical rx36 running HP-UX 11.23 on theirs internal disks.

The first server is on my datacenter and I make a DRD clone on a SAN disk which is replicated (EMC SRDF) on my second datacenter.
(The second server is on the second datacenter)

I try to reboot my second server on the SAN "replicated" disk, but there is a LVM error at boot. Actually, as the SAN HW path (/dev/dsk/cxtydz) are not identical between the 2 datacenters, the file /etc/lvmtab is incorrect on the replicated disk : it containts the /dev/dsk/cxtydz of the source server.

If I boot in LVM maintenance mode, I can vgexport/vgimport the vg00, then I can correct the boot and root vol and finally I can boot my server (without problem) on the SAN disk, but it's a manual correction.

My question is : On my target server, when I'm running HP-UX on the internal disks, how can I change the file /etc/lvmtab of the "replicated" DRD disk ? (I can vgimport the vg00clone and mount the root vol)
Is there a "magic" command to rewrite the target lvmtab ? (If any, I can write C code, but I don't know the lvmtab structure)

Thank You.

PS :Of course, replication is stopped for the test.

[smatador]

Hi,
I never do that, so I give some idea.
First you could drd mount the replicated disk and try to modify in it the ioinit conf file. With 11.31 you could modify online the io path, but here you have a drd disk on 11.23...
I think also that you have to change some file like bootconf.
Another idea, what's happen when you delete lvmtab on the drd and after you try to boot to the second server ?
HTH

[Kapil Jha]

I think, once you have fixed this, it would work. The only thing you want to do everytime is chose which disk u gonna boot the server from, you have to set it with setboot.

Provided that you are local and EMC boot disk are used exclusively.


BR,
Kapil+

[suncoast]

Hi Lucas,

Not sure it this is still an issue. However, you can use the vgscan command to rebuild the /etc/lvmtab file. You can use the -p option to run in preview mode to see that it will do before actually doing it. Also, I recommend making a copy of your existing /etc/lvmtab file prior to rebuilding it with the vgscan command.

Singram

[Kapil Jha]

I don think there is any way to do it automatically, even if you try to boot your server with DRD disk u wont be able to boot(never tried though) becuause DRD environment you activate via command.

In destination server device names would be different for sure.
So if you do it for the first time vgimport you can reboot it n number of times without doing anything...till ur replication is stopped.

Once replication started you would have to manually do it.

BR,
Kapil+