[OpenAFS] rh9 + openafs 1.2.10

Nick Burrett nick@dsvr.net
Tue, 18 Nov 2003 09:30:11 +0000


David Thompson wrote:
> We're trying to deploy Red Hat 9.  I'm running kernel 2.4.20-20.9, openafs 
> 1.2.10 (_not_ Derek's RPMs - our own build from the openafs.org sources), and 
> I'm running into some strange failure modes.  What I'm seeing is random 
> processes hanging in the kernel (and never coming back), and/or unreasonably 
> long times reading data out of /afs.  It's not deterministic; I can run some 
> random program many times, and then once in a while it hangs in the kernel, 
> and all I can do is 'kill -9' the process.
> 
> Is this new?  Is anyone else seeing it?  Is there a fix somewhere in CVS?

I'm also seeing this, though I can't kill the processes.  I thought it 
might be because I'm accessing /afs though a different location set by:

$ mount --bind /afs /mount/foo/afs

I'm using Derek's RPMs on a RedHat 9 server.  The client is a vanilla 
linux-2.4.22 machine with Derek's RPMs re-compiled to target that kernel.


Regards,


Nick.



-- 
Nick Burrett
Network Engineer, Designer Servers Ltd.   http://www.dsvr.co.uk