Write Failed message with 9.1-RC3

[Mark Saad]

On Nov 26, 2012, at 4:01 PM, "Reed A. Cartwright" <cartwright at asu.edu> wrote:

> I'm new to this list...
> 
> I'm running a bioinformatics server using 9.1-RC3 (64 cores, 512GB
> ram).  I have a ZFS raid-z2 array attached to an LSI controller with a
> SSD cache drive.

Can you tell us more about this server . Can you post the output of pciconf -lv ,
What do you have in loader.conf and sysctl.conf ? Do you se a custom kernel build;  What are the changes you made to that config ? 
> 
> Since upgrading to 9.1-RC2/3 (for AVX support), I have been
> experiencing hard drive lockups with the message "write failed"
> printed to the console.  After this reading from the hard drives no
> longer work, but the machine is not locked up.  If the appropriate
> files are in the cache, I can log in and execute programs.  I know the
> LSI driver has been updated in 9.1 and I have updated my cards' bios
> to match.  It doesn't seem to make a difference.
> 
> Once I was able to run top, and saw that many processes were stuck in
> the 'tx->tx' state.

What does your zfs config look like can you post the output if zpool status ? What zfz options have you set on pool and filesystens ?

Lastly In a shutdown -h now , does the system properly shutdown ,does it crash and dump core ?
> 
> So far, no corruption appears to have occurred in the drives.
> 
> I'm about to downgrade to 9.0, but I wanted to know if anyone has any
> idea what the issue is.
> 
> --
> Reed A. Cartwright, PhD
> Assistant Professor of Genomics, Evolution, and Bioinformatics
> School of Life Sciences
> Center for Evolutionary Medicine and Informatics
> The Biodesign Institute
> Arizona State University
> _______________________________________________
> freebsd-stable at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
> To unsubscribe, send any mail to "freebsd-stable-unsubscribe at freebsd.org"

---
Mark saad | mark.saad at longcount.org