Holger Parplies <wbppc <at> parplies.de> writes:
...
>
> On the other hand, on the computers where it matters (servers, BackupPC), RAID
> 1 has been running for years without a real problem (I *have* seen RAID
> members
> dropped from an array without understandable reasons, but, mostly, re-adding
> them simply worked; more importantly, there was no interruption of service).
In a prior life, I used Linux software RAID 1 (mirroring) in a startup. We were
dumb and poor, and built servers using mail-order parts, including hard drives.
For interesting reading, lookup raid vibration error. At the time, this was a
sort of new phenomena, and even the disk manufacturers didn't understand why
there were very high failure rates of hard drives in RAID configurations. Most
consumer systems had 1 hard drive, and most "enterprise" systems used SCSI for
the higher performance. Both worked fine.
But when companies started using vanilla IDE hard drives in RAID configurations
at very busy sites with a lot of concurrent IO, things broke. We had a failure
rate of about 10% on Maxtor hard drives, with behavior that has been mentioned
here: the drive fails with a read error, is taken out of the array, but can be
re-added with no problem. Then later, the drive fails again, at a different
address. Repeat until you have no more hair to pull out. :)
The problem is that when 2 or more drives are in close proximity and are
accessed concurrently, drive 1 can be sitting over a track waiting for a sector
to rotate around, and a seek in drive 2 can knock drive 1 "off track". This
leads to a read error in drive 1. If it is part of a RAID, the usual OS
response is to fail the drive since it is mirrored. If the drive were NOT
mirrored, the OS usually just retries the seek+read, and is successful. So a
RAID1 configuration makes this configuration break more often.
The drive makers soon realized what was happening and came out with "enterprise
class" IDE drives. These have a feature to monitor and adjust the head position
following a seek to ensure the head stays "on track".
Once we switched to SCSI drives, this problem went away.
Jim
--
HashBackup: easy Unix onsite/offsite backup
www.hashbackup.com
------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network
management toolset available today. Delivers lowest initial
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd
_______________________________________________
BackupPC-users mailing list
BackupPC-users AT lists.sourceforge DOT net
List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki: http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
|
