Re: ORA-03137: TTC PROTOCOL INTERNAL ERROR :  .... as any body ever see this?
Date: Fri, 3 Oct 2008 10:01:52 -0700 (PDT)
Hi, we are seeing the same ORA-03137:TTC error on our RAC nodes and corresponding
javax.servlet.ServletException:  java.sql.SQLException: No more data to read from socket
errors on our tomcat 5 app servers using the thin client We too have had a case open with Oracle - but with the random unreproducible nature of the issue we have not been able to progress very far with them
I've taken tcpdumps of the inter-connect and its perfectly clean - now I am looking at the bonded (RedHat) VIP interfaces...
thanks for any feedback,
On 31 Aug, 23:56, Cristian Cudizio <cristian.cudi..._at_yahoo.it> wrote:
> On 29 Ago, 18:47, joel garry <joel-ga..._at_home.com> wrote:
> > On Aug 29, 2:33 am, Cristian Cudizio <cristian.cudi..._at_yahoo.it>
> > wrote:
> > > Our customer is having sometimes this error .
> > > ORA-03137:TTCPROTOCOL INTERNAL ERROR :      
> > >  
> > > On Oracle 220.127.116.11 on AIX 5.3
> > I have no idea what I'm talking about, but are you seeing lots of
> > things like FIN_WAIT_2 when you do a netstat -a?
> > I've off and on had a completely different problem on different
> > hardware, that turned out to be a flaky server that was being talked
> > to as a background printer spooler process, which would cut off
> > communication incorrectly, which would take an hour for the tcp layer
> > to figure out it should kill. This would take away resources from
> > other processes, with the net effect of random socket errors in the
> > client/server environment, made much worse by an abusive transaction
> > monitor.
> > Perhaps nothing to do with your problem, but such things need to be
> > considered with random socket errors.
> > It could also be a bug or OS patch or... I notice metalink lists a
> > linux bug 7148457, "closed, could not reproduce" with the  
> >  ... so up your SR severity, that makes me think Oracle isn't
> > handling some situation properly, getting the same error parameters on
> > different platforms, even if the root cause is configuration or non-
> > Oracle bugs, or even the drivers.
> > jg
> > --
> > @home.com is bogus.
> > What must the secrets be? http://www.signonsandiego.com/uniontrib/20080829/news_1b29qualcomm.html
> Thanks Joel for tips. We hadn't considered the possibility of network/
> socket problems. We have
> to investigate more on this. With netstat -a there not a lot of
> FIN_WAIT_2 but i must check
> when servers are more loaded. With netstat i'va only noticed that on
> AIX on
> /etc/services there are this two lines:
> ncube-lm 1521/tcp # nCube License
> ncube-lm 1521/udp # nCube License
> while having oracle listener on port 1521.
> I've also to put up the SR serverity to make Oracle Support handling
> the request.
> Cristian Cudizio
Received on Fri Oct 03 2008 - 12:01:52 CDT