isw device for volume broken after opensuse livecd boot

Hall, Eric R eric.r.hall at
Fri Sep 28 00:09:56 UTC 2007

>-----Original Message-----
>From: ataraid-list-bounces at [mailto:ataraid-list-
>bounces at] On Behalf Of Tiago Freitas
>Sent: Thursday, September 27, 2007 2:59 PM
>To: Fang, Ying
>Cc: ATARAID (eg, Promise Fasttrak, Highpoint 370) related discussions
>Subject: Re: isw device for volume broken after opensuse livecd boot
>The third disk was already in but not part of the array, and I chose
>it to rebuild in OROM bios. As I never did a rebuild before (the pc is
>two weeks old and this is my first raid) I don't know how the process
>is done.
You might want to read some of the documentation on RAID.  Digg had a
link to the "Best RAID Tutorial, Ever!" which is a good place to start.
That, or Wikipedia.

>Should I have chosen one of the array members to rebuild?
Generally not.  In a RAID configuration, if there is a failed disk, you
don't want to try to rebuild to one of the existing/failed members since
it's bound to fail again.  That's why you either have a hot spare or add
new hw to rebuild to.  However, your situation is a bit different.  The
problem is that you've tried so many permutations that it's like trying
to work against quick sand: the more you fight, the deeper in it you go.
We're still trying to figure out how the metadata on sdb (PVF904...VDN)
managed to have sda added and removed several times.

>I have all the data backed up in an external disk, I just wanted to
>help fix the bug in OROM because it's weird and may happen again. But
>I can't wait too much time because I will need the pc up and running
It's good to know that you have the data backed up.  We've spent most of
the day trying to figure out how to produce the issue so that you could
recover the data.  However, if this isn't a necessity, then we won't
spend as much time on trying to figure out how to recover the data.
We're having a difficult time reproducing the issue all together.  Don't
wait on us; rebuild your system.

>The size difference is really weird too because both disks were
>manufactured in Jully 2007 and have the exact same model...maybe
>that's what triggered the failure?
No, the drives them selves have unique serial numbers and those are what
device-mapper and dmraid rely upon to determine which disk is who.
Hard disk drives that are produced in the same time frame and same model
might be tested differently and depending on the outcome of the test the
manufacturer might dictate its maximum size differently.  The disks
might have come from different manufacturing plants or have firmware
that reports its total size in bytes differently.

>I'm gonna test the disks in Hitachi Fitness tool to be sure they are
You don't have to contact the list, but could you let me know if the
disks are 'ok' so we might have some closure on this oddity?



More information about the Ataraid-list mailing list