hot pluggable drives

[DonS_1]

I have a server with two hot pluggable drives in a mirrored raid config. Whenever we make a change that may affect the OS, we pull one of the drives as a 'safety net'without shutting down first. We have never had any trouble using the 'safety net' drive to boot up to when our changes have gone bad. My question is, is this practice ok, or should we knock all users off then shut down the server, then pull the drive. Does it make a difference either way? Recently a co-worker raised this as a possible concern. Thanks for your response!

[Igor Karasik]

Two years ago I asked similar question and received answer from HP.
The answer was like:
"The disks are hot pluggable,
and not hot-UNpluggable, so in order to unplug disk - do shutdown. Only if disk failed and controller detected it you can replace disk without shutdown"

[Marko H.]

You have to shut down first.

If the server is running and you pull the disk out the controller detects that one physical disk failed and he waits for a physical disk to rebuild logical drives.

Do shut down first. Then at POST you'll get the message that one physical disk is missing and two options:
- to rebuild when new disk is inserted
- not to rebuild when new disk is inserted.
Pick the one you wish, make your changes.
If anything goes wrong with your changes you have your "safety net" that is working for sure.

[Colin_29]

The only issue may be with warranty. I was once told by a Compaq specialist that the drives record, on board, their number of hot swaps and that any more than three will invalidate the warranty.

This kind of makes sense because, especially with the higher rpm drives, when you remove one completely without giving it time to spin down, you can damage the bearings. They have a real gyroscopic effect, we were shown this on a course once.

Since then I have heard contradicting reports about warranty replacements, and HP won't tell you because they don't want to encourage you to break drives