Ataraid-list Digest, Vol 42, Issue 15

[Hall, Eric R]

>-----Original Message-----
>From: Matthias Koenig [mailto:mkoenig at suse.de]
>Sent: Thursday, August 23, 2007 3:27 AM
>To: ATARAID (eg, Promise Fasttrak, Highpoint 370) related discussions
>Cc: Fang, Ying; Hall, Eric R
>Subject: Re: Ataraid-list Digest, Vol 42, Issue 15
>
> As I was told by Eric this should be ISW and
>not DDF1, I guess it is not possible to let the configurator on
S5000VSA
>board create DDF1 format?
Nope.

> So, I am wondering why this RAID is detected
>as DDF1, if it should be ISW. But this will have to answered by the
>reporter (I do not have this hardware so I cannot do tests personally).
My bet is that the disks were installed in a system that had another
controller that used DDF1.  Or, the submitter is using an add-in card on
the Sapello system which is what has written the metadata on the disks.

As Ying says, metadata is disbursed at the end of the drive...  Can you
get the submitter to run a couple of commands to see if there is any ISW
metadata?
# cat /sys/block/sdb/size
Subtract two from the returned numeric value
# dd if=/dev/sdb skip=<returned number minus two> count=2 bs=512|
hexdump -C
Look for something like "Intel Raid ISM".  If that text is there, then
the disks have Intel ISW metadata.
If that doesn't return anything, keep looking through the end of the
disk to see where some metadata like info might be hiding:
# dd if=/dev/sdb skip=<returned number minus 2000> bs=512 | hexdump -C
And look for some human readable text.


>Currently I have the output of dmraid -s -ccc available:
But, is there a way for you to get the dmraid -n and lspci output from
the submitter?  This way we can see what hardware is present and the
contents of the metadata.