> Recently, our cell (catt.ncsu.edu) installed a copy of AFS 3.4 GA onto our
> DECstations and our Sparc fileserver. Our clients are DECstation 2100s and
> 5000/{25,200}s. The 2100s have 24 Mb of RAM, and the 5000s have between 24-32
> Mb of RAM. The server is a Sparc 5 with 32 Mb of memory; all of the AFS
> server processes run on this system.
> The Sparc seems to run fine with AFS 3.4 GA + Solaris 2.4. Our AFS clients
> are not faring so well. The one AFS 3.4 ULTRIX client we are running, a
> DECstation 5000/200, is crashing intermittently with this message:
> ---
> cpu 0 panic: getdcache
[... stuff deleted ...]
> Something similar happened with our ULTRIX systems under 3.4 beta.
> Another interesting note: we've noticed that some files are getting
> corrupted in the cache.
> Please send possible solutions, if any. If there are no immediate solutions,
> then we will be regressing back to AFS 3.3a for ULTRIX.
> --
> s. parkerson <[EMAIL PROTECTED]> "...one of the perks of living in a
> another CSC undergrad at NCSU free society is the inalienable right
> head sysadmin, CATT Program at NCSU to debase ourselves in a wide variety
> http://www.catt.ncsu.edu/~witsend of ways." (Sam & Max, Freelance Police)
We reported this to Transarc a month or more ago and they said they
were working on it. I haven't heard anything since.
One important thing to note- AFS 3.4GA (and 3.4beta) both cause cache
corruption (on the Ultrix boxes)- if you switch back to 3.3a your
cache will still be corrupted. For us, the cache is a separate
partition so I just booted single-user and newfs'ed the
/usr/vice/cache partition. If your cache lives on a partition with
other data you'll need to boot single-user and "rm -rf *" in
/usr/vice/cache.
Good luck!
Kevin Hildebrand
University of Maryland, College Park
Project Glue Systems Developer
