Bug #81472 for DBD-DB2: Prepare statement errors showing up after upgrade

Tue Nov 27 03:34:04 2012 Morten.Bjornsvik [...] evry.com - Ticket created

Subject:	Prepare statement errors showing up after upgrade
Date:	Tue, 27 Nov 2012 09:29:19 +0100
To:	"bug-DBD-DB2 [...] rt.cpan.org" <bug-DBD-DB2 [...] rt.cpan.org>
From:	Bjørnsvik Morten <Morten.Bjornsvik [...] evry.com>

Hi All I upgraded our middleware to: mod_perl2 -- 2.000007 DBD::DB2 -- 1.85 DBI -- 1.616 perl 5.16.1 DB2 v9.5.7 (but seen on 9.7 also) If running through mod_perl/apache I get the following 2-3 lines a minute: [ERROR] 2012/11/27 01:06:20 /opt/s3/server/S3DB.pm:1732:- psql() - unable to prepare sql:"select name,value from config", errstr():[IBM][CLI Driver] SQL30081N A communication error has been detected. Communication protocol being used: "TCP/IP". Communication API being used: "SOCKETS". Location where the error was detected: "192.168.1.28". Communication function detecting the error: "send". Protocol specific error code(s): "32", "*", "0". SQLSTATE=08001 , err():-30081 at /opt/s3/server/Const.pm line 240. If we use the same modules directly there is no problems. We have our own sth cacher We only connect and create and never finish() until we exits or the cache is filled up. we cache on process and sql txt, so there is no sharing between processes. If i use $DBH->connect_cached() without our cacher, apache works almost fine, but then the daemon starts complaining big time. Makes me believe there are some issues with how DBI manages the connections. how it disconnects. Have anyone else seen this errors? Any idea how to debug this? Thanks -- Morten Bjoernsvik Senior Developer Secana Evry Cardservices AS Nydalsveien 28 0484 Oslo +47 92 44 83 02 EDBErgoGroup is now Evry

Wed Nov 28 05:46:02 2012 Morten.Bjornsvik [...] evry.com - Correspondence added

Subject:	[rt.cpan.org #81472]
Date:	Wed, 28 Nov 2012 11:45:27 +0100
To:	"bug-DBD-DB2 [...] rt.cpan.org" <bug-DBD-DB2 [...] rt.cpan.org>
From:	Bjørnsvik Morten <Morten.Bjornsvik [...] evry.com>

Hi I do not care about the log messages, but from time to time I get a hang that need to wait for the databasehandler to time out. This seem to be related to mod_perl 2.0.7 If I install mod_perl 2.0.5 I do not get any hangs, but I'm unable to compile mod_perl 2.0.5 on perl 5.16.x But I see the same with perl 5.10.1 with mod_perl 2.0.7 but it works with mod_perl 2.0.5 -- MortenB

Fri Nov 30 23:24:32 2012 opendev [...] us.ibm.com - Correspondence added

Hi Morten, I exactly would not be able to tell what the problem could be. First you might have to check what has changed in mod_perl, might be we will get some hints from there. Second thing which could be happening is might be your apache process or thread got a end signal and it called the close connection call. But now you had a active statement handle, in other thread, which your are trying to use and hence the communication error. We can also quickly diagnose and check what could be happening by getting CLI traces. Could you get the CLI trace for your application run? http://publib.boulder.ibm.com/infocenter/db2luw/v9r7/topic/com.ibm.db2.luw.admin.trb.doc/doc/t0054958.html Thanks Praveen

Fri Nov 30 23:24:33 2012 The RT System itself - Status changed from 'new' to 'open'