Congratulations :D

Out of curiosity, how did you fix the homehost issue?

On Sun, Nov 15, 2009 at 12:45 AM, Leslie Rhorer [off-list ref] wrote:

quoted

quoted

    Should I stop the arraya nd re-assemble it with -vvv?

    Also, what about my question about the bitmap?  Should I delete it,

despite

quoted

the fact mdadm says there isn't one, or should I attempt to add it

again, or

quoted

should I wait until I do the checkarrray?

There is nothing to delete.  Just try to add it again.
Though I would really like to know why it hangs.  If that happens
again, see if alt-sysrq-T shows anything useful.

       Well, it hasn't hung again, but now that I can catch my breath I can
provide a little detail.  As I mentioned before, the original hang occurred
when I first issued:

`mdadm -G /dev/md0 -b internal --bitmap-chunk=65535`

       Mdadm complained about the bitmap chunk being an incorrect size.  I
don't know if it actually created a bitmap anyway using the default size, or
what, but I assumed the error meant the bitmap was not created.  Going on
this assumption, I immediately issued the command again, changing only the
bitmap chunk size:

`mdadm -G /dev/md0 -b internal --bitmap-chunk=65536`

       That's when it hung.  I tried stopping the array, but I couldn't
umount the file system because it was dirty.  Any commands to mdadm - even
informational ones and even

`cat /proc/mdadm`

       would cause the xterm to hang.  Even a shutdown would not work.  I
had to shut down  the server with the power switch.

       When the system booted, mdadm would not assemble the array.  I think
Neil has figured out why.  Subsequent attempts to assemble the array also
failed, at least in part because /dev/md0 had been created without the array
being active.  I deleted /dev/md0, but subsequent failed assembly attempts
again created the file.  Finally I deleted the file once again and then
assembled using

`mdadm -A /dev/sd[a-j]`

       and it worked, although the array was marked as dirty, and the first
attempt to access it caused a resync.  When the resync was done, I fixed the
"homehost" issue Neil had spotted and corrected /etc/mdadm/mdadm.conf to
show all 10 drives.  Finally, I added back the bitmap, and now everything
seems fine.  I rebooted the server, and it comes up with all 10 drives in an
assembled array with an internal bitmap.

--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
       Majed B.
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html