Coda File System

Re: coda client hangs

From: Jan Harkes <jaharkes_at_cs.cmu.edu>
Date: Tue, 31 May 2005 16:55:13 -0400
On Tue, May 31, 2005 at 01:37:25PM -0600, Patrick Walsh wrote:
> On Tue, 2005-05-31 at 14:15 -0400, Jan Harkes wrote: 
> > I've been working on the iterators over the weekend. The 'safe deletion'
> > parts are removed from the core iterator functions and I think I got all
> > places where we were doing unsafe deletion.
> 
> Is there a patch I should apply to 6.0.10?  

Currently the patch is still only in my working directory, I want to
test the servers still, but won't get to that until tonights weekly
backups have completed.

I also don't know how much this patch depends on other changes that are
already in CVS, since it touches a lot of things all over the place. So
I may end up snapshotting CVS as a new release if the servers seem to be
working with the new iterators.

> Hmmmm.  It never occurred to me that giving tokens to root might cause a
> problem... 

Coda used to disallow tokens for root. Technically it shouldn't be a
difference, you would hit the same bug with a regular user if the token
renewal happens concurrently with something like weak-reintegration or
accessing backup volumes. It is because we have a periodic server probe
as a root user which made it more likely that you hit the problem when
clog refreshed the tokens for root.

> 	Is it odd that though the servers are supposed to be checked every five
> minutes, there was no check at 00:05 or 00:25?

We only probe the servers if we haven't heard from them in a while. The
servers also send backprobe messages. So if we either recently made an
RPC call, or if the server sent it's probe first we skip the probing.

Jan
Received on 2005-05-31 16:56:36