Iam back again :)

[Sprint Unix Team]

When a primary root disk fails..will the system still be up and running with the mirrored root disk OR
the system reboots and comes up with the mirror..

[Craig Rants]

Yes a system should be up with a mirror. Remember mirror disk and alternate boot disk are not necessarily synonymous! You can specify an alternate boot disk if you want that is not your mirror disk.

[G. Vrijhoeven]

Hi,

The system will keep up and running.

[A. Clay Stephenson]

Well, that depends. If you have mirrored all the lvol's on the disk then your box will never know anything bad has happened other than the system monitoring daemons telling you. If you also made both disks bootable and allowed booting without quorum you can reboot
with little intervention. If you have hot-plug or hot-swap disks you can even replace the disk and never stop the server.

Clay

[Eileen Millen]

It think this is one of those, it depends questions.
I got this from HP
Start at
http://itrc.hp.com

Click on Maintenance/Support (on the left hand side)
Click on Search Technical Knowledge Base
Login to the ITRC.

Search String: RCFAX* mirror


Hope it is still there.
There are some common misconceptions at the bottom.
Eileen

[Uday_S_Ankolekar]

Hi,
System will up and running as I know..
But it 's better to have a script to stale check for logical volume and inform admins.

-USA..

[Darrell Allen]

If you mirror all logical volumes on the root disk the system keeps running. As far as rebooting, you have to set up the "mirror disk" as a boot disk to be able to boot from it regardless of whether or not you've mirrored all the LVs on it.

Darrell

[Sprint Unix Team]

You are all right..But its a long list of replies..do u guys all want points assigned
;-)

[Eileen Millen · Accepted solution]

The attachement is a message that I got from
HP support.

Eileen

[A. Clay Stephenson]

Hi again:

I think I will try to put the 'does the box stay up?' issue in perspective. Understand, that I do mirror each lvol including swap in vg00. I also have the boot strings set to override quorum. Autosearch and Autoboot are enabled. I have enough servers so that in any given year about 6 boot drive failures occur.

Unless the internal drives are hot-plug, I always boot from external drives which are hot-pluggable. The mirrored drives are always on a separate SCSI bus in a separate enclosure with separate power. All of this is basic high-availabilty stuff. Given all those condition, in over 5 years and again with several boot drive failures per year, I have NEVER had a machine go down because of drive failure and have never had to shutdown the server to replace the drives. I keep spare drives on hand
and generally have a boot drives replaced and resynched in just a few minutes.

Food for thought, Clay