tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Problem with raidframe: bad disk or bad memory, how to tell?

        Well.  Checking each partition individually  adds mystery to the
problem.  Each of the disks reports no errors, and the fsck'ing of each
disk individually looks clean.  Yet, running fsck on /dev/rraid0a yields
similar results to the ones I already posted.  My guess is that I'm seeing
memory corruption.  This makes me wonder  if I need to reboot the machine
without syncing the disks first.  Perhaps I need to install memtest86 to
see if I can exercise the memory and find out where the problem is.

On Jul 10,  8:39pm, Manuel Bouyer wrote:
} Subject: Re: Problem with raidframe: bad disk or bad memory, how to tell?
} On Thu, Jul 10, 2008 at 11:30:18AM -0700, Brian Buhrow wrote:
} >     Hello.  I have a NetBSD-3.1 system in production which seems to be
} > experiencing a filesystem issue on its raid1 root partition.  The symptoms
} > are that when we tried to run a backup last night, the dump program never
} > completed, and it created an endless dump file.  Also, running fsck on the
} > root partition yields inconsistent results, especially when you consider
} > that this machine isn't that active.
} >     I think that either one of the disks is silently corrupting data,
} > or I have some bad memory in the box.  I'm almost certain it's a hardware
} > issue, but the question is, how to tell which hardware specifically is the
} > problem?  I could try failing one of the disks, but then, if I pick the
} > wrong one, I'm really in trouble.  Is there some utility I can use to
} > compare the contents of the two disks to see if they're equal?  the box is
} > a production box, as I mentioned, so I'd like to avoid bouncing it up and
} > down, at least until I've figured out what  the problem is most likely to
} > b.
} >     Any thoughts?
} For each disk you can create a 4.2BSD partition which matches the RAID-1
} partition, and so run fsck on each drive independantly.
} The matching 4.2BSD partition usually starts 64 sectors after the RAID
} partition
} -- 
} Manuel Bouyer <>
}      NetBSD: 26 ans d'experience feront toujours la difference
} --
>-- End of excerpt from Manuel Bouyer

Home | Main Index | Thread Index | Old Index