On Wed, 3 Feb 1999 13:43:35 -0500 , John Zoltak <[log in to unmask]>
wrote:

>Bill,
>
>I have a 957RX running on the Nike array, HP3550A. You can create a ldev
>1 out of a mirrored pair RAID 1, and the rest can be groups of RAID 5.
>This has worked very well for us. Recently we had a drive fail. It
>worked perfectly. I got hold of a replacement, swapped the drives and
>about 3 hours later we were redundant again. Everything was working the
>whole time. You can pick up a used Nike array with drives in the $10K to
>$20K region.
>

Buy yourself one more disk mech as a hot spare and you can be
redundant even BEFORE you get a chance to swap out the failed mech.

Can anyone confirm or deny that MPE/iX 6.0 supports the auto-trespass
feature (which in the case of an array with two storage processors,
allows all physical disks to be accessed by only one SP in case the
other fails. Right now the admin must go into the disk array and force
this to happen during failures, but HP-UX systems, for instance, can
be configured to do this automatically. It's really the last barrier
to MPE/iX having a really bulletproof HP disk array solution.......)
---

Mark Landin                   "For anyone who was never good at
T. D. Williamson, Inc.         anything, technology has been a
UNIX Sys. Admin                real boon" --- my mom