> -----Original Message-----
> From: Mladen Turk [mailto:[EMAIL PROTECTED]]
> Sent: 18. lipanj 2002 19:39
> To: 'Tomcat Developers List'
> Cc: 'GOMEZ Henri'
> Subject: RE: jk 1.2 freese/snap time in JTC ?
>
>
>
>
> > -----Original Message-----
> > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> > Sent: 18. lipanj 2002 19:26
> > To: Tomcat Developers List
> > Cc: GOMEZ Henri
> > Subject: RE: jk 1.2 freese/snap time in JTC ?
> >
> >
> > I'm not sure I understand how request_rec gets recycled while
> > we are in map_uri_to_worker - it is supposed to happen after
> > the request completes. It's very weird - maybe there's
> > another problem. Can we avoid a mutex in the critical path ?
> >
> > Does the same problem exists in jk2 ?
> >
>
> I've tried that with the Enter/Leave CriticalSection, but as
> I done more
> testing, I still occasionally get 1 or two handler
> "jakarta-servlet" not
> found, returning server error.
>
> I'll try to dig into that (at least I think I'm on the right track).
> The strangest thing that map_uri_to_worker occasionally
> returns NULL for
> the same uri.
>
Can't believe it!
We have a race condition beetwen the:
Alias /examples /tomcat/webapps/examples
and
JkMount /*.jsp ajp13
If I comment out the Alias directive there is no more colision, and if I
slow down the map_uri_to_worker putting the Sleep(100), then for the
each request there is 'not found handler'.
Any ideas how to solve that?
MT.
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
