james-server-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Noel J. Bergman" <n...@devtech.com>
Subject RE: james 2.1.3: deadlock in mordred code
Date Wed, 14 Apr 2004 18:23:26 GMT
> > > I am about to commit a patch to CVS.  Would you please check it?
> > > did you send it to me? Can't find it!
> > Was testing before posting.  Should be there now.
> Looks fine to me. This patch surely stops the JAMES being brought down.

Actually, I didn't test enough.  Just fixed again.  The sleep call was in
the block, and needed to be outside of it.  That caused James to run, but
slowly.

> a hard limit of 2 minutes seems very short to me

As I said, with a hard limit of 1 minute, I have seen 32 incidents in 3
years.  I'm reluctant to leave all of those resources open on a DB server,
which has a limited number of them, if there is a programming error, but I'm
curious to see now what people report.

> at the initial insert by the smtp handler the other peer
> should be signaled the error condition

Yes, an exception caught by the spooler when saving the message will be
reported in the protocol response.

> it stays in the spool even when for example inserting it
> into an error mailrepository fails

If there is an exception during the pipeline, it should stay in the spool.

> I added [the second if-statement] later. But obviously before submission.

Ah, got it.  :-)

	--- Noel



---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
For additional commands, e-mail: server-dev-help@james.apache.org


Mime
View raw message