[email protected] wrote:
> Skylar Thompson:
>   
>>> According to your config.gz, you set CONFIG_AUFS=y so you don't need
>>> the sec_perm patch. And the ksize patch is unnecessary too. It is for
>>> linux-2.6.21 and earlier.
>>>       
>> Does it hurt to have it included? If it does, I can rebuild the kernel
>> and test again.
>>     
>
> Basically nothing hurt. I just thought people doesn't like apply any
> patch to the vanilla kernel. If there is something hurt, a few bytes of
> memory are consumed to export those symbols.
>
>
>   
>> Examining this process actually led me to stumble onto what I think is
>> the real problem, which might be unrelated to aufs. It looks like the
>> udev tmpfs that is started in the initramfs isn't being migrated
>> properly to the new root filesystem. This means that init creates
>> initctl over NFS in the r/w aufs branch, and appears to be causing some
>> locking problems. I don't know if the locking problems are aufs-related
>> or not, but obviously the solution is to get /dev migrated properly.
>>     
>
> According to your Magic SysRq output, aufs requested the read operation
> to NFS and waited for the completion. I guess it is /dev/initctl which
> is a named pipe and no process writes to it. Finally the read operation
> never complete.
>
>
>   
>>> And I don't think CONFIG_AUFS_EXPORT=y, CONFIG_AUFS_ROBR=y,
>>> CONFIG_AUFS_SHWH=y and CONFIG_AUFS_RR_SQUASHFS=y are necessary for you.
>>>       
>> Does it hurt to have these enabled? I wasn't sure which ones I needed,
>> so I enabled all the ones that looked like they could be useful.
>>     
>
> It will hurt the perfomance of speed and memory usage.
>
>
>   
>> Thanks for the suggestions!
>>     
>
> You are welcome.
> Anyway I will be able to sleep well and celebratea new year. :-)
>
>   

Sorry for the delay, but I just wanted to let you guys know that I think
we've solved the aufs problem. We had been clearing out the
node-specific directories (the read-write branch) after each use. For
some reason NFS was serving up the stale filehandles, but refusing to
lock them. Once we stopped clearing out the read-write branch, our
problems went away. We didn't need to do that, so it's no big deal.

Thanks for the help again, and thanks for a great product. Our work
really would be a lot harder without aufs.

-- 
-- Skylar Thompson ([email protected])
-- http://www.cs.earlham.edu/~skylar/


------------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It is the best place to buy or sell services for
just about anything Open Source.
http://p.sf.net/sfu/Xq1LFB

Reply via email to