Hi Mandy,
Does this mean you have been working on breaking up the xfs extent data
array into small chunks? I always wanted to do that, just never had the
time.
Does this make kmem_realloc go away? It is not strictly correct in that
if kmem_realloc fails, it still frees the original memory. XFS is rather
dead at that point, but in theory the error could have been rippled up
the call chain and xfs left with the old extent array.
Steve
p.s. I see you got married, congratulations.
Mandy Miklos wrote:
>>Christoph,
>>
>>I'm writing to find out the status of the SGI - XFS bug #345
>>http://oss.sgi.com/bugzilla/show_bug.cgi?id=345
>>
>>I am experiencing similar issues on a client's server that runs SUSE
>>9.2with kernel
>>2.6.8-24.14-bigsmp.
>>
>>We are getting several errors that show up in /var/log/messages, which are
>>similar to those of the previous posters. In our case, there is no
>>peer-to-peer file sharing taking place. This is a bioscience cluster at a
>>local University, which mounts a 4TB RAID array over NFS.
>>
>>I'm curious to know if this issue has been fixed in other, more recent,
>>kernels... or if it is still outstanding.
>
>
> Aaron,
> The fix for this issue is in the final stages of testing. Look for it
> to be pushed out late Jan to early Feb '06.
>
> Mandy
>
|