RE: BROKEN_PIPE from Weblogic

2003-10-09 Thread DENNIS WILLIAMS
Agreed. By the way, since my original posting, this error has occurred
several more times.
   Yes, it is a catch-all error that sounds a lot like an ORA-3113. Since
they using the Oracle thin JDBC driver, which talks directly to the
listener, Net8 isn't involved. If the server is shut down or crashes,
apparently this error is also returned, like an ORA-3113. But the server
hasn't crashed or shut down. 
   This is a catch-all error and BEA has a long list of things that can
cause this error. The database can cause this error under a few
circumstances. Like the database closed the connection. Naturally the
developers turn to the DBA at that point.
   At this point I just want to know what to say when they ask did the
Oracle server close the connection? In my years of working with Oracle I
haven't found it to spontaneously close connections. But how can I prove
that Oracle isn't closing the connection? Can anyone think of a log that
would be written to if it did?
   Should the developers be capturing more information in their Java code? I
can recall back in the days of novice C programmers writing Oracle database
calls that new programmers would not check error conditions and if an error
occurred the program would continue merrily along, processing garbage.
Apparently some of our other Java programs are coded to retry the connection
if this occurs.

Dennis Williams
DBA, 80%OCP, 100% DBA
Lifetouch, Inc.
[EMAIL PROTECTED] 


-Original Message-
Sent: Thursday, October 09, 2003 7:41 AM
To: Multiple recipients of list ORACLE-L


do they mean the ora -03113 end of file communication error(number may be
off)? Alot of stuff causes that error. 

that is a very vague response on their part.  
 
 From: DENNIS WILLIAMS [EMAIL PROTECTED]
 Date: 2003/10/08 Wed PM 11:34:24 EDT
 To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
 Subject: RE: BROKEN_PIPE during Weblogic J2EE deployment
 
 Apparently BROKEN_PIPE is a pretty generic error. Quite a few problems can
 cause this error. One problem BEA mentions is if the database closes the
 connection. Does anyone have a suggestion for what I should reply to
this?
 Are there any situations under which Oracle would close a connection?
Would
 that event appear in any log?
 
 Dennis Williams
 DBA, 80%OCP, 100% DBA
 Lifetouch, Inc.
 [EMAIL PROTECTED] 
 
 
 -Original Message-
 Sent: Monday, October 06, 2003 1:34 PM
 To: Multiple recipients of list ORACLE-L
 
 
 As Weblogic is not an Oracle product, it is not surprising that Metalink
 returned nothing.  Try your query on Google.
 
 Ian MacGregor
 Stanford Linear Accelerator Center
 [EMAIL PROTECTED]
 
 
 -Original Message-
 Sent: Monday, October 06, 2003 10:19 AM
 To: Multiple recipients of list ORACLE-L
 
 
 Our web administrator was deploying a new J2EE application to production
and
 received the error broken_pipe. We are using the thin client. Apparently
 Weblogic tries to verify existence the tables the application uses. The
 deployment was otherwise successful in that the error occurred several
times
 when deployment was retried but testing the app worked fine. They feel it
 was a database error, but I don't see anything on my end. Naturally they
are
 nervous about this error just going into a critical new application.
 Apparently if the database was down, this is the error returned on the
 Weblogic side, but the database was fine and I couldn't find any errors. I
 searched for broken_pipe on Metalink and the search returned empty. Has
 anyone had any experience with this Weblogic error?
 
 Dennis Williams
 DBA, 80%OCP, 100% DBA
 Lifetouch, Inc.
 [EMAIL PROTECTED] 
 -- 
 Please see the official ORACLE-L FAQ: http://www.orafaq.net
 -- 
 Author: DENNIS WILLIAMS
   INET: [EMAIL PROTECTED]
 
 Fat City Network Services-- 858-538-5051 http://www.fatcity.com
 San Diego, California-- Mailing list and web hosting services
 -
 To REMOVE yourself from this mailing list, send an E-Mail message
 to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the
 message BODY, include a line containing: UNSUB ORACLE-L (or the name of
 mailing list you want to be removed from).  You may also send the HELP
 command for other information (like subscribing).
 -- 
 Please see the official ORACLE-L FAQ: http://www.orafaq.net
 -- 
 Author: MacGregor, Ian A.
   INET: [EMAIL PROTECTED]
 
 Fat City Network Services-- 858-538-5051 http://www.fatcity.com
 San Diego, California-- Mailing list and web hosting services
 -
 To REMOVE yourself from this mailing list, send an E-Mail message
 to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
 the message BODY, include a line containing: UNSUB ORACLE-L
 (or the name of mailing list you want to be removed from).  You may
 also send the HELP command for other information (like subscribing).
 -- 
 Please see 

RE: BROKEN_PIPE from Weblogic

2003-10-09 Thread Jeff Herrick


I don't think it's accurate to say that NET8 isn't involved when
you're talking directly to the listener =8-)  More correctly
you are eliminating the NET8 client DLL's...but you could still
have the connections dropping due to transient network failures.

I have a particular problem with CISCO switches and TRU64 Alpha
boxes in my environment. JDBC thick connections drop but because they
are 'thick' I get a SQLNET.LOG file on the client that shows
a TNS-12203 (unable to connect to destination) error. Other apps
like Maximo report an ORA-3113 but also show the TNS-12203 error
in their sqlnet.log files. When I do a level 16 listener trace I can
see the connection being closed (read() error) but I've never been
able to figure out why  =8-(

The only thing that moderates the problem is to make sure that the
Alpha is set to 100MB/Full Duplex instead of auto-detect as well as
hard-coding these values on the switch port. Ditto for the NT app
servers.  friends don't let friends auto-negotiate  =8-)

But I still see periodic TNS-12203's just less of them

Try doing a level 16 trace from the listner end.

Jeff Herrick

On Thu, 9 Oct 2003, DENNIS WILLIAMS wrote:

 Agreed. By the way, since my original posting, this error has occurred
 several more times.
Yes, it is a catch-all error that sounds a lot like an ORA-3113. Since
 they using the Oracle thin JDBC driver, which talks directly to the
 listener, Net8 isn't involved. If the server is shut down or crashes,
 apparently this error is also returned, like an ORA-3113. But the server
 hasn't crashed or shut down.

[snip]

-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.net
-- 
Author: Jeff Herrick
  INET: [EMAIL PROTECTED]

Fat City Network Services-- 858-538-5051 http://www.fatcity.com
San Diego, California-- Mailing list and web hosting services
-
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).