[OpenAFS] Re: afs stalled for large files, opeanfs 1.6.1, ubuntu 12.04, particular network

Andrew Deason adeason@sinenomine.net
Sun, 20 Apr 2014 22:52:43 -0500

On Fri, 18 Apr 2014 18:11:59 -0400
Jeffrey Hutzelman <jhutz@cmu.edu> wrote:

> This sounds like an MTU problem -- your connection to the network in
> question includes a segment with a lower MTU than those to which the
> fileserver and client are directly connected.  Perhaps there is a VPN or
> other tunnel in between?
> Try running this command (as root) on the client:
> ip link set dev eth0 mtu 1200
> Replace "eth0" with the name of the relevant interface.

You can also pass -rxmaxmtu 1200 to afsd, to only affect the openafs
client. The actual number would need to be a little smaller than 1200 to
be analogous to the command Jeff gives, but that doesn't much matter;
just try using different values.

Also, thanks for the extra debugging information, but the fstrace logs
and such probably aren't needed in this particular case (the above info
should be enough). But for future reference, some comments on them:

It's helpful to know in situations like this if we're hanging on copying
data out of AFS or into AFS. So, you could have separate tests for
copying a file from /afs/.../ to /tmp/, and from /tmp/ to /afs/.../, to
see which part of the operation we're hanging on.

And the fstrace logs provided don't look... complete (they cover a very
short period of time). You need to generate some more AFS activity
before killing the fstrace command (just do that until you see the
output file change), or you could probably unbuffer the output with a
command like 'stdbuf'. fstrace meanwhile should probably be less stupid
and flush its output.

But again, that's all just for future reference; you don't need to
re-run those commands to provide that info again, since the extra info
is probably not needed.

Andrew Deason