On Fri, 7 Apr 1995, David Gadbois wrote:
> I have a 2100 running OSF/1 3.0 with 4 RZ28-VAs hooked to a SWXCR-EB in
> a RAID 0 configuration. On two occasions, the SWXCR has reported read
> errors once each on two different drives in the group. Both times the
> SWXCR marked the drive as failed and took itself offline. Both times I
> powered the system down, pulled the failed drive, reseated it, and used
> the SWXCR configuration software to mark the failed drive as optimal.
> The first time I did this the filesystem on the SWXCR mounted, and I was
> able to do a level 0 vdump of the whole thing. I am in the process of
> doing a second level 0 vdump after the second go-round.
>
> Now, the drives are only a couple of months old and have been running
> fine for the several weeks I have had the system. It seems awfully
> suspicious that lightning would strike twice like this. So, my current
> assumption is that something else is wrong, maybe with the SWXCR, maybe
> with the SCSI cabling, maybe with the internal StorageWorks shelf the
> drives are on. Has anyone seen similar failures and their solutions?
> Or suggest potential problems and ways to troubleshoot them? My DEC CE
> is not, alas, familiar with the StorageWorks setups, so I am afraid I am
> in for a component swap-fest (with its attendant hassles and wastes of
> time) unless I can narrow the problem down for him.
>
> Thanks,
> --David Gadbois
>
Concerning RAID on the 2100,
We have the same basic configuration: 2100 4/200, 1cpu, 256 MB Ram, 2 2Gbyte
rz28 drives configured as ADVFS and 8 Gbytes Level 5 RAID (5 rz28 storage-
works drives in shelf utilizing the SWXCR-EB 3 channel RAID built-in RAID
controller). The system including the ADVFS has not given me any
problems, but the RAID, which was being used for some NFS performance test-
ing, dropped out twice with hardware failure messages. In my case I had no
real data out there so I used the Raid Config Utility to reinitialize the
RAID array, and then everything was fine!!. Presently, I would not recommend
utilizing the RAID array for user disk space due to this. I am hoping to
here more about this from this mail list. By the way I stopped the NFS
testing for other reasons, so it is not as if the problem has gone away.
Todd
Received on Fri Apr 07 1995 - 07:25:10 NZST