Data backup

[Jeff Ferris]

Good morning. Need insight or direction where I can find the best information for setting up a backup in the event that if a drive would fail all proccesses/data would automatically route to another drive(s) seamlessly to continue operation.

Thank you.
Jeff

HP/UX 11.0
Sybase 11.5

P.S. This thread has been moved from Disk to  HP-UX > sysadmin. -HP Forum Moderator

[Andreas Voss]

Hi,

for your purposes the best way is
mirroring all the disk data with MirrorDisk/UX

Of course this depends on how much disks you have and also the license for mirroring.

Regards

Andrew

[Rick Garland]

Mirror-UX would be the way to go for disks. But if you are also concerned with the entire system failing on you, look at MC/ServiceGuard as well. I getting the idea you want to minimize the single points of failure.

[Anthony deRito]

If you have a "NIKE" disk array you can configure the LUN so that you will have a primary mirror part and secondary mirror part. This is called hardware mirroring. You do not need to purchase additional software if you use hardware mirroring. If the primary mirror disk fails, the secondary mirror disk will be fully functional. When you hot swap the bad disk, the mirror copy will rebuild.

Tony

[Tim Malnati]

Some form of RAID protection is the only reliable form of protection for disk failures. The MirrorDisk product is what I would call the entry level form of raid protection on an HP box. Basically it is a simple software addition combined with additional attached disks that are fully mirrored (raid 1) with the primary drives. If a failure occurs on any single drive, the opposite drive in the configuration handles i/o until such time that it can be replaced. There are all sorts of different hardware arrays available that can perform raid protection at the hardware level (although there is always some level of software running somewhere, be it on a array controller or whatever). In general these arrays are more expensive, but can be more cost effective in larger disk storage situations. These usually have additional capabilities for load balancing, etc, that often can offset potential costs associated with platform duplication and the like.