LDAP delay when connected DC goes down

[Jeremy Allison]

On Wed, Jan 12, 2005 at 06:08:17PM +0100, Alex de Vaal wrote:
> 
> On Tue 11 Jan 2005 at 21:31, Jeremy Allison wrote:
> 
> >> On Tue, Jan 11, 2005 at 11:08:49AM -0800, Joe Meadows wrote:
> 
> >> Thanks for your patch.  My partner Alex tested it and it looks like it 
> >> does take care of the problem.  CUPS is still a problem as there is a 2 
> >> minute delay when printing from a Windows machine before the client 
> >> finally comes back with an error message.  I haven't looked into this 
> >> yet so maybe his logs will reveal some important clues.  Just mentioning 
> >> it in case there are some obvious or well known reasons why this is 
> >> happening.
> 
> > Can you do an strace/gdb attach at this point so we know where in
> > the code we're hanging ?
> 
> I'm the "test partner in crime" of Joe Meadows ;)
> 
> I've sent you an "off-list" e-mail with a log level 10 of winbindd and a
> document with
> timestamps of the steps I took in the test (which corresponds with the time
> stamps in
> the winbindd.log)
> 
> Hope this is helpful in finding the CUPS problem when the primary configured
> DC server
> is down and winbindd reattached to the domain via the secondary configured
> DC server.

I don't have it yet I'm afraid. Also - I may need more than this. If it
hangs then I may need to see a gdb backtrace of winbindd at the point you're
waiting for a reply.

Jeremy.