System restores during disaster recovery

[Mike Tufariello]

Hi all,

For disaster recovery testing and real situations, I bring up my DR system using my Ignite tape, which includes all of vg00. I then restore my full backup tape, which also includes vg00 mount points.

In a DR test the full backup tape would be older than my Ignite tape. Should I not restore any of the vg00 mount points from my full backup tape because they are older than the Ignite tape?

During a real DR, where the Ignite tape is likely to be older than the full backup tape should I restore any of the vg00 mount points from the full backup tape? And if so should I use the overwrite option?

I use Data Protector to back up and restore data.

Thanks very much,

Mike

[Ivan Krastev]

Better way is to keep your Ignite tape up to date. You can take Ignite backups once a week, or after every system update/patching.

If your vg00 hold only system files and no applications, the difference will be only /var/log/* ,/tmp/* and /var/tmp/*, but they doesnt affect system operation.

regards,
ivan

[Steven E. Protter]

Shalom Mike,

A real recovery is going to normally involve two steps.

Ignite restore of vg00 and vg information for the other volume groups.

Data restore.

No real data should be included in vg00 and therefore the Ignite tape.

SEP

[VK2COT]

Hello everyone,

If your servers run HP-UX 11i v2 or v3,
then you could think about Dynamic Root Disk too. It is cloning for the O/S. As long as you can use some extra disks, DRD is
a nice way to do it.

Look how simple it is. I just did it
on an HP-UX 11.31 server
for my students (I am a Senior Unix
Instructor in HP Education and,
coincindentally, working on updating the
training materials for H3064S and H3065S
courses worldwide):

# drd clone -t /dev/disk/disk8 -x overwrite=true

======= 07/02/08 13:09:41 EST BEGIN Clone System Image (user=root)
(jobid=myserv)

* Reading Current System Information
* Selecting System Image To Clone
* Selecting Target Disk
* Selecting Volume Manager For New System Image
* Analyzing For System Image Cloning
* Creating New File Systems
* Copying File Systems To New System Image
* Making New System Image Bootable
* Unmounting New System Image Clone

======= 07/02/08 13:42:57 EST END Clone System Image succeeded. (user=root)
(jobid=myserv)

As well, you might even think about
creating Ignite DVDs. Check
/opt/ignite/data/scripts/examples/make_opticaldisc_recovery if you have a decent version of Ignite.

Cheers,

VK2COT

[Sp4admin]

Like Ivan,

We run a ignite backup weekly. We just wrote a script to run one a week.

sp,

[Mike Tufariello]

Thanks to all for the replies.
However, no really answered the questions.

To provide a little more information. I use Data Protector to backup my Unix system. The DP database resides on vg00 in /var/opt/omni. I do make an Ignite tape weekly and when I do I stop DP services to make sure the Ignite tape includes the DP database.

If I recover my system using an Ignite tape from Sunday, the DP database will include all backup and restore sessions to that point, making recovery easy. However, if I recover the system using backup tapes created after the Ignite tape was created it will be more difficult.

Nevertheless, my Ignite tape has the following mount points /, /stand, /var, /usr, /tmp, /opt from vg00. My backup tapes also include these mount points.

Should I overlay the mount points restored from the Ingite tape with the mount points that will be restored from the backup tape ? Or just not restore those mount points at all becasue they were restored from the Ignite tape?

Thanks,

[Pete Randall]

Other than /tmp and /var/tmp, the mount points you mention should be nearly completely static unless you swinstall a product in /opt, at which time I would suggest doing another Ignite tape anyway, so I think your question is really moot as it shouldn't make any difference.


Pete

[Rita C Workman]

What Stephen mentions is basically what we do here for DR. As was said, vg00 doesn't change that often, by keeping your ignite tapes current - your covered.

The only adjustment is the backup server that runs Data Protector. For rehearsal, you could create a tarfile of /var/opt/omni & /etc/opt/omni and then drop that on a tape and copy it back at your DR site on your corresponding DR server.
Better still for a real DR - create the tarfile daily (dated) and ftp the it to a small server at the DR site. That way in the event of a real disaster you already have it there to restore once we rebuild the box. Once you have your tapes there - your ready to restore data.

Just a thought,
Rgrds,
Rita