dead records in locking.tdb (was: Locking database cleanup?)

Dhruva Krishnamurthy dhruvakm at gmail.com
Sun May 21 04:10:48 GMT 2006


Hi,
 Just an idea with my limited knowledge of the core internals.
- Use the aproach suggested by Jeremy, clean up if no process is
running with the PID in the TDB records
- For every forked smbd process (on each new client connection), have
it clean up the TDB's with it's PID (child process's PID) at startup.
This makes sure that no stale records can affect that process's
working.

Handling this in a two phases might minimize stale records remaining
from non-existant processes.

with best regards,
dk

On 5/20/06, Jeremy Allison <jra at samba.org> wrote:
> On Fri, May 19, 2006 at 11:28:50PM -0600, Steve Williams wrote:
>
> > In this case, the log files might reflect the process died, but the KILL
> > signal finds a process.. how could that be?  It was someone's shell that
> > logged in afterwards, and by coincidence got the same PID as the process
> > that terminated abnormally.
>
> Reusing PID's very quickly will break lots of apps, not just
> Samba. Just something else I dislike about the *BSD's....
>
> Jeremy.
>


-- 
dhruva (dk)
Views expressed in this email are my personal views unless explicitly mentioned.


More information about the samba-technical mailing list