Coda File System

Re: Strange timeouts / no such device

From: Martin Ginkel <mginkel_at_mpi-magdeburg.mpg.de>
Date: Fri, 23 Mar 2007 04:50:41 -0400
Stephen J. Turnbull wrote:
> The UI for repair could be made smoother, I'm sure, but I don't think
> it's possible to reduce it much below
> 
> cfs beginrepair file-in-conflict
> discardalllocal
> cfs endrepair
> 
> which is what we already have (at least in theory, repair was pretty
> buggy the last few times I've tried it).  What would be nice would be
> a better interface for more complex repair situations, where multiple
> changes to the same file need to be integrated, or where it's not
> obvious which changes should be discarded.

But this is only the theory. In practice for me this fails in at
least 30% of the cases. especially for server-server conflicts, where
you have no alternative to repair.

In local-server conflicts the crazy workaround is to checkpoint the cml,
reinit venus and replay the tar-file with the files afterwards.
Usually I give repair 1 try, if this fails due to
* venus crash
* VIOXXXX errors

I go on with reinit. I'm definitely faster with that version.
Although, to my impression the frequency of bogus-conflicts
has decreased.

	Martin
Received on 2007-03-23 18:39:53