[CentOS] mdadm quandary

Arun Tagarsi tiger at tigerclan.co.uk
Sun Oct 2 12:10:05 UTC 2005

      I wonder if has seen a problem of the sorts I am describing as at the
moment it is driving me nuts.

We have an array of LSI disks (Hardware Raid 5)  which are attached to a
Linux node. The Node has a LSI929 dual fibre channel card that has a channel
attached to controller A and B of the array.

So we have a list of devices under /proc/partitions  mapping what we are
expecting to see , for example the devices

/dev/sdd1 (A)  and /dev/sdx1 (B ) would form a raid 0 device under linux to
make a file system slightly under 1TB.

This file system was created using the older mkraid utils at the time but
has been in production for over a

This node and the array unfortunately had a power outage and the resulting
problem faced by us is the following.

The raid start up fails with as it cannot find a device /dev/sdai1 . This
device is not part of the /etc/raidtab but is a normal device defined in the
/dev directory.

There are no references to the /dev/sdai1 device in /proc/partitions at all
. All devices expected as /dev/sdd1 and /dev/sdx1 are listed and show the
expected sizes. so I presume these are the only devices that the Linux box
would be able to work with.

If I run a mdadm --examine of the /dev/sdd1 it shows me that indeed
/dev/sdd1 and it is lun 0 of the MD device but it also has references to the
/dev/sdai1 device.

The mdadm --examine /dev/sdx1 shows up as  /dev/sdai1 and is lun 1 of the MD

This is where I am totally confused and I hope I was able to explain the
problem .

Is there a way to change the super block from referencing sdai1 to sdx1?
without compromising the array . The data on the array we suspect is intact
and would take a great deal of time to restore.

Thanks for your time.


No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.344 / Virus Database: 267.11.9/116 - Release Date: 9/30/2005

More information about the CentOS mailing list