[omniORB] Troubleshooting a server that is becoming inaccessible during the night.
Michael L Bendickson
bendi003@tc.umn.edu
Thu, 9 Nov 2000 09:55:29 -0600 (CST)
The port is dynamic. I'm letting the ORB pick it.
There should not be another instance of my server. There is another
non-omniORB TCP server on the system, that is VERY busy. Could that server
be grabbing the port, or would this only occur with another instance of my
server?
I wrote a small TCP server with a fixed port on my NT box with winsock, and
tried to start it twice. bind() fails as I would hope. Are there situations
in NT where this would succeed?
I'll check the state and state-reason of the threads the next time this
happens.
Thanks,
-Mike
> Do you select a fixed port number for the server? If so, could there be
> another instance of the server started using the same port number? On unix,
> the other server instance will just fail to grab the port. On NT, I think
> the OS just let the new server instance to grab the port and needless to
> say this cause great pain to track down the problem.
>
> Next time when it happens, could you look at the state of each thread in
> the process? It is important to know what the rendezvous thread is doing.
> I don't think the thread has gone away, or you would have seen a message in
> stderr as you have already turned up the trace level.
>
> Sai-Lai
>
> >>>>> Mike Bendickson writes:
>
> > Ok, it happened again today. I am not able to connect to that port using
> > telnet when this happens. I was able to connect with telnet ealier and saw
> > a "GIOP" message before getting kicked off.
>
> > So, something has trashed the socket then, or the thread is not sitting on
> > accept?
>
> > Is there something else I should look for? Has anyone else experienced this
> > with 2.8.0 on NT?
>
>
> --
> Sai-Lai Lo S.Lo@uk.research.att.com
> AT&T Laboratories Cambridge WWW: http://www.uk.research.att.com
> 24a Trumpington Street Tel: +44 1223 343000
> Cambridge CB2 1QA Fax: +44 1223 313542
> ENGLAND