Good Evening Oscar I am Glad to hear you were able to work out the
synchronisation issues
This is very strange error with Oracle..I dont understand why cache-ing
anything from Oracle would help
as a sequence as to be incremented for every insert access
*Perhaps we can discuss this offline*
Saludos Cordiales desde "Boston Con Nieve"(Cordial Salutations from Boston with
Snow)
Martin ______________________________________________ Porfavor no interrumpir
esta communicacion..GraciasFrom: oscas...@hotmail.com
To: java-user@axis.apache.org
Subject: RE: Axis2 not access to db after several requests
Date: Thu, 7 Feb 2013 13:44:32 +0100
Hi all:
About this error, we´ve discover something about it. The error has no
relation to Axis2.
We´ve found that Oracle 11 sometimes (even Oracle doesn´t know when and
why) block access to sequences (it seems lile when several threads request
value of the same sequence, sequence is like blocked.
Oracle has advised to turn on a cache on sequence on Oracle 11, and they
hope it´ll solve problem.
Thanks as allways and i only want to say what we discovered
Kind regards,
Oscar
From: mgai...@hotmail.com
To: java-user@axis.apache.org
Subject: RE: Axis2 not access to db after several requests
Date: Tue, 18 Dec 2012 12:21:44 -0500
since this is a production issue we would rather know the exact error..
in this way we could offer the one solution which when tested will remedy this
exact problem
Saludos,
Martin
______________________________________________
Porfavor..no altere o disrupta esta communicacion...Gracias
From: oscas...@hotmail.com
To: java-user@axis.apache.org
Subject: RE: Axis2 not access to db after several requests
Date: Tue, 18 Dec 2012 15:50:29 +0100
Hi Marty
I´ll try to do all you have written and get a good log
Thanks
Regards
From: mgai...@hotmail.com
To: java-user@axis.apache.org
Subject: RE: Axis2 not access to db after several requests
Date: Mon, 17 Dec 2012 08:04:33 -0500
From: oscas...@hotmail.com
To: java-user@axis.apache.org
Subject: Axis2 not access to db after several requests
Date: Mon, 17 Dec 2012 13:38:47 +0100
Hi all,MG>Hello Oscar
Since several weeks ago i am having pretty big problems with my Ws on
Axis2 1.6.2 i´ll try yo explain in few words:
I have a Ws deployed on Axis2 that access an Oracle database in order to
retrieve info.
I built a jar which encapsulates access to databse, i mean i create a jar
in which i implement a connection pool.
To make this connection pool i used DBCP 1.4. My server is Tomcat 6 and i
use Java 6 on HP-UXMG>standard implementation
After making several requests to my Ws it starts to fail. I put several
logs in order to know step by step when it fails and i found that it fails the
first time a request need to access databaseMG>possibly your connection URL and
or Database connection parameters are incorrectMG>be careful to record your
maxActive and maxIdle connections and track maxActive/maxIdle connections
combinationsMG> http://commons.apache.org/dbcp/configuration.html and this
only happens after severral request (sometimes 20, sometimesm 8000 and
sometimes two weeks after)MG>possibly the connections are not being reclaimed
MG>8000 maxActive is alot of active Connections for the connectionPool to
handle you may want to throttle that backMG>which driverClassName are you
implementingMG>also be mindful that ReadOnly should be enabled if you have
ReadOnly connections
MG>how is your transaction commit and rollback executed..are you waiting too
long for commit or rollback and possibly starving other Threads of their
connection request
I thought it was connection pool problem (maybe some dbcp issue) so i
change my jar and start to use the framework BoneCP which i found works
faster.MG>not familiar with this product ..does anyone have knowledge of this
product to help Oscar
I got same problem, do anyone have any idea? Because i´m becoming crazy, it
fails also in Pro environment and i have to watch it almost 24 hours because
it´s an important wsMG>log everything and send us the logsMG>
http://axis.apache.org/axis/java/reference.html#AxisLoggingConfiguration
Any idea will be welcomed
ThanksMG>your welcome...please send us the logs..so we can better ascertain
the cause of this problem
Kind regards
