Matt Macander wrote:
> Damien Hull wrote:
>
>>Matt Macander wrote:
>>
>>
>>
>>>We have a 1 TB Terastation from Buffalo Tech with a RAID 5 array, that
>>>will no longer boot after a botched firmware upgrade.
>>>
>>>The four 250 GB hard drives appear to be fine and I think that they are
>>>a standard Linux software RAID array (formatted with XFS). I am trying
>>>to mount the RAID array on a different computer so I can pull some data
>>>off before sending the unit back to get replaced.
>>>
>>>I tried hooking them up to another computer and booting into the latest
>>>version of Knoppix. Unfortunately, the full capacity of the drives
>>>doesn't seem to be available and the RAID array cannot be reconstructed...
>>>
>>>hde: max request size 128 KiB
>>>hde: 488397168 sectors (250059 MB) w/ 2048 KiB cache, CHS=30401/255/63
>>>attempt to access beyond end of device
>>>hde: rw=0, want=1460886144, limit=486962282
>>>I/O error in filesystem ("hde3") meta-data deve hde3 blok
>>>("XFS_read_buf") error 5 buf count 4096
>>>...
>>>and so on for drives hdf, hdg, and hdh
>>>
>>>
>>>
>>>I could find the root partition hde1 from Knoppix and the raid config
>>>file mdadm.conf:
>>>ARRAY /dev/md1 level=raid5 num-devices=4 UUID=a39b74nb:...
>>>devices=/dev/hda3,/dev/hdc3,/dev/hde3,/dev/hdg3
>>>
>>>I think the problem may be the very large size of the drives, though I'm
>>>not sure how to get around it. Any ideas?
>>>
>>>Matt
>>>---------
>>>To unsubscribe, send email to <aklug-request@aklug.org>
>>>with 'unsubscribe' in the message body.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>Are you sure this is software raid? I have to ask. What follows is only
>>true if you have software raid.
>>
>>Does knoppix have the raid tools. Fedora uses mdadm to manage software
>>RAID. You will also need the raid device in /dev. On my server I have
>>/dev/md0 for /boot and /dev/md1 for everything else.
>>
>>Also check to see if Knoppix can read XFS. Not all Linux distributions
>>can read XFS.
>>---------
>>To unsubscribe, send email to <aklug-request@aklug.org>
>>with 'unsubscribe' in the message body.
>>
>>
> Definitely software raid (according to the manufacturer, and I also
> found the mdadm.conf file).
>
> I know Knoppix can read XFS becasue the one partition that I can read
> off the four drives is the non-raid XFS partition on the first drive
> (where the /etc folder is located).
>
> I tried adjusting the mdadm.conf file and running mdadm --assemble,
> but it could not find superblocks on any of the devices.
>
> That's when I noticed the errors in the log when the partitions were
> initially scanned during bootup:
>
>hde: max request size 128 KiB
>hde: 488397168 sectors (250059 MB) w/ 2048 KiB cache, CHS=30401/255/63
>attempt to access beyond end of device
>hde: rw=0, want=1460886144, limit=486962282
>I/O error in filesystem ("hde3") meta-data dev hde3 block
>("XFS_read_buf") error 5 buf count 4096
>
>That makes me think the full drive capacity is not being recognized somehow...
>
>
>
>
>
>
You may have the same problem I had when my RAID stopped working. I'm
still not sure what happened but my guess is that the file system became
corrupt on one of the drives (RAID 1). It wouldn't allow me to rebuild
the array. I was able to backup all data and reinstall Fedora. I believe
I had the same superblock problem as well.
You may have one bad drive or one drive with a corrupt file system.
Taking out that drive may allow you to rebuild the array. This is only a
guess though.
Hope you get your data back.
For those that are reading this on the list RAID is not a replacement
for BACKUPS. They can and do fail.
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Wed May 11 16:35:57 2005
This archive was generated by hypermail 2.1.8 : Wed May 11 2005 - 16:35:57 AKDT