From: Dirk Kleinhesselink (dkleinh@phy.ucsf.edu)
Date: Tue Mar 29 2005 - 20:02:25 EST
Thanks to Colin Bull, Tarmo Linnamaegi and Bluejay Adametz for their
replies: Essentially to prevent the automatic replacement of a failed
disk in a raidset on a HSG80 controller, use the command:
SET {RAIDSET_NAME} NOPOLICY
The bus with the failed disk should be quiesced prior to physically
removing the disk - I did some googling as my documentation for my HSG80
did not have a good description about the buttons on the HSG80 and I found
a nice document from HP (don't have a URL for it anymore - it was a
customer advisory with the description: HSx80 Support for Device Removal
Replacement 'Hot Swap') which described the procedure: the HSG80 back
(where the scsi connectors hookup) has 6 small square lit buttons, one for
each bus. In my case, bus 1 and 3 had failed disks and the buttons for
those busses were lit. To quiesce a bus, press the button for about 2
seconds and about 10 seconds later, all disks on that bus (shelf) will
flash their amber/red lights. When the lights flash, the bus is quiesced
and the disk can be removed.
I replaced the disk and then gave the commands:
SET {RAIDSET_NAME} REPLACE=DISKYYYYY
SET {RAIDSET_NAME} POLICY=BEST_PERFORMANCE
and the raidsets began reconstruction.
-- Original message below ----
My alpha cluster connects to raidsets on a HSG80 controller via SAN. Last
week, first one, then another disk failed on it -- fortunately, both
failed disks were in different raidsets. I unfortunately don't have a hot
spare and so the raidsets have been running reduced until the replacement
disks -- and I'm going to get a spare disk this time -- arrive. What I'd
like to know is how can I manually replace each failed disk so that each
disk goes into the correct raidset ? Also what is the procedure (and
console comands) for replacing a disk ?
Thank you very much for any help.
Dirk
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:50:17 EDT