LDEV's To Replicate With Continuous Access (CA)

[Andy Torres]

We recently implemented CA from our local XP10K to a remote XP10K array. We'll need to adjust our DR plans accordingly, but one question remains for me: Do we replicate our boot devices (HP-UX 11.11)? Can they actually be viable boot devices to *different* server hardware? In our HP BCS agreement we can never guarantee same server hardware.

Has anyone had experience recovering their replicated HP-UX systems at the remote site with different server hardware? For example, we have rp4440's here, but we might have to use a Superdome partition for recovery - whatever BCS provides at the time of the recovery.

If vg00 isn't a viable boot device on the replicated array we'll have to keep Ignite in the plan for initial recovery and do vgimports for the data areas.

I'm interested in hearing some of the different methods used in similar circumstances as well. Thanks for the info in advance.

[Sundar_7 · Accepted solution]

You dont duplicate the boot device. The common approach is to use CA for the data replication and use Ignite-UX tapes for vg00 recovery.

If the BCS agreement cannot gurantee the same server hardware type, I am afraid you may run in to lot of unforeseen issues.

This could even impact the ability to meet your SLA to recover the systems, if you had to rebuild the OS from the scratch at the recovery site.

[Steven E. Protter]

Shalom,

I have some experience.

make_tape_recovery/make_net_recovery will recover vg00 in the same hardware line. e.g. rp5450 to rp5470 for example.

Going beyond that is pushing it.

Golden images are more flexible, but increases recovery time on site.

The only sane way to deal with data, especially database data is to have a cold backup to recover to followed by log recovery if possible.

SEP

[Andy Torres]

We've been through a DR exercise now and I just thought I'd update this a little.

We ended up replicating vg00 and successfully used the secondary volume to boot on remote servers provided by HP at their Alpharetta recovery center where our remote XP is located. After some dodgy SAN switch zoning, etc., our remote DR team was able to boot from the vg00 secondary volume (we cautiously use only one 18GB LUN for vg00) and recover/mount the data volumes on other volume groups.

We understand that this may or may not work perfectly if we get a server very different from the rp4440's we're using now, so we are also getting IGNITE tapes twice a week on them in case we have to recover vg00 manually. That is also not a solid as the suggested golden images, but for us it fits our recovery parameters.

Just thought I'd update the subject. I'll leave this one open for a little while longer to allow further discussion, then close it. Thanks again for the input. Much appreciated.

[Andy Torres]

All done with this one. Thanks again for the input.