Coda File System

Re: Proposing an improved release schedule

From: Jan Harkes <jaharkes_at_cs.cmu.edu>
Date: Mon, 13 Aug 2007 22:53:58 -0400
On Mon, Aug 13, 2007 at 06:49:49PM -0400, Greg Troxel wrote:
> > Possibly, although most systems seem to use the identical errno values,
> > considering how 'well' Coda has worked without doing any translation.
> > The error numbers that were moved around were those that do not exist on
> > typical UNIX system (EINCONS - conflict detected).
> 
> maybe something else.  it was flaky again today.  BTW, I don't get why
> my machine was even trying to talk to testserver - I reinited to avoid
> that!

After a reinit I wouldn't expect it to pick out testserver all by
itself. There really shouldn't be anything referencing that realm.

I did add the functionality to gcodacon to remove volumes from the list
when venus drops them from the cache. This of course pretty much
immediately identified volumes that were not deleted, every time I ran
cfs fl the reference count went up by one.. Turns out that when we
initially mount a volume we incremented the refcount twice instead of
only once.

Interesting how even a simple visualization of what is going on inside
the client really makes some bugs easier to find.

Jan
Received on 2007-08-13 22:56:40