[Linux-cachefs] more NFS mounting hangups with CacheFS

[Steve Dickson]

Jacob L. wrote:
> Hello,
> 
>    Much like the issues Chris had I have run into the same. I am using fc3 all the latest packages updated through yum. My kernel was prepared the same way as Steve suggested:
> "I started with the linux-2.6.11 tarball then applied the patch-2.6.12-rc3.bz2 
> patch and the 2.6.12-rc3-mm3.bz patch". The util-linux patches were then forced 
> from http://people.redhat.com/steved/cachefs/util-linux/. I have setup the cache 
> block device, initialized it, all that goes well until NFS hits the equation. 
> As soon as I try to mount NFS with the fsc option it just hangs there, cant kill
> the process, cant reboot the machine, same issues. Has there been any progress 
> towards a solution? 
I would be good to see where the mount is hanging...
use "echo t > /proc/sysrq-trigger" to dump a system
trace to /var/log/messages. That tell where the hang
is occurring.

> Can we expect this to be fixed in the new version David announced a few days ago?
Well these are the "next generation" patches. Meaning these will be
replacing the current patches in the -mm and the ones that we will
be porting to FC/RHEL kernels... So it might make sense to migrate
to them.

> 
> I tried to load this on another system, Core 2, a few weeks back and was able to 
> mount the nfs with the fsc option. As soon as I initiated any type of traffic to 
> the mount (10megs of web traffic from the frontend, CachFS machine, to the 
> backend via NFS mounted with fsc) the system load shot up to the 100's and 
> the machine came crashing down.
Again, having the oops output would be helpful....

steved.