So, I'll try to answer this sequentially :)
1 - I've tried reverting back to the previous kernel. No good.
2- This is a stock redhat kernel - 2.6.18-194.26.1.el5 to be precise.
I was previously using 2.6.18-194.el5.
RAID drivers afaik are modular in redhat. (sata_mv being the one in
question.)
I've done some digging and it's been suggested to zero out the superblock
on the drives and then recreate the RAID. Does anyone have experience in
doing this or advice to whether or not it will wipe my data out as well?
Thanks!
Adam
On 12/14/2010 12:55 PM, Brad Bendily wrote:
You may want to figure out which kernel you were on before and go back
to that one, just
to see if it's actually the kernel. If i you revert, and it works ok,
then you can read
what was changed in the newer kernel docs. I'm assuming you're using
your distro's kernel and
not rolling your own? If so, you can wait for your distro to upgrade
to a newer kernel that fixes your
problem, or roll your own kernel.
I had a situation once on a Dell 2650 a certain kernel from our distro
caused high I/O wait times.
This was fixed by a new kernel from the distro. Fortunately I didn't
have to roll my own. I may have
tried that, but I didn't realize the kernel was the problem until I
had upgraded.
It could possibly also be updates in the mdadm software. But, for
troubleshooting purposes
I would only change one thing at time at this point. Starting with the kernel.
Also, is your raid drivers compiled in the kernel, or as modules?
You could possibly just revert the raid driver, but those may want a
matching kernel.
Or you could compile your own newer raid drivers against your current kernel.
bb
On Tue, Dec 14, 2010 at 12:35 PM, Adam Yates<[email protected]> wrote:
Hi all;
I'm having a few problems with a fileserver at work and would like to bounce
some ideas off
anyone that could help. I'm ayates83 on #brlug on irc.freenode.net (does
anyone still inhabit the room?)
Email is fine too.
Basic setup is this:
2 sets raid-6. 22 drives in each set plus a hot spare.
After I upgraded a kernel, the drives randomly reassign letters under the
raid and cause
the set to break on boot. I can use mdadm --assemble --scan -fv to
forcefully bring the
raid back online but it crashes soon thereafter.
Any input is appreciated!
Thanks,
Adam
--
Adam Yates
Systems Administrator -- Research Infrastructure
Center for Computation and Technology
232 Johnston Hall,
Baton Rouge, LA 70803
W: 225.578.8235 C: 225.663.0218
<[email protected]>
_______________________________________________
General mailing list
[email protected]
http://brlug.net/mailman/listinfo/general_brlug.net
--
Adam Yates
Systems Administrator -- Research Infrastructure
Center for Computation and Technology
232 Johnston Hall,
Baton Rouge, LA 70803
W: 225.578.8235 C: 225.663.0218
<[email protected]>
_______________________________________________
General mailing list
[email protected]
http://brlug.net/mailman/listinfo/general_brlug.net
