Hello Robert,

thanks for your quick reply.

This is what the full_cluster.log states after having started the
application:

DEBUG controller.core.Controller Connection accepted for 
e178044177.adsl.alicedsl.de
DEBUG controller.core.Controller Starting new ControllerWorkerThread
DEBUG controller.core.Controller Connection from 
e178044177.adsl.alicedsl.de/85.178.44.177:51076
DEBUG controller.core.Controller Starting new VirtualDatabaseWorkerThread
DEBUG controller.core.Controller Connection accepted for 
e178044177.adsl.alicedsl.de
DEBUG controller.core.Controller Starting new ControllerWorkerThread
DEBUG controller.core.Controller Connection from 
e178044177.adsl.alicedsl.de/85.178.44.177:51077
DEBUG controller.core.Controller Connection accepted for 
e178044177.adsl.alicedsl.de
DEBUG controller.core.Controller Connection from 
e178044177.adsl.alicedsl.de/85.178.44.177:51078
DEBUG controller.core.Controller Starting new ControllerWorkerThread
DEBUG controller.core.Controller Starting new VirtualDatabaseWorkerThread
DEBUG controller.core.Controller Starting new VirtualDatabaseWorkerThread
DEBUG controller.core.Controller Connection accepted for 
e178044177.adsl.alicedsl.de
DEBUG controller.core.Controller Connection accepted for 
e178044177.adsl.alicedsl.de
DEBUG controller.core.Controller Connection from 
e178044177.adsl.alicedsl.de/85.178.44.177:51080
DEBUG controller.core.Controller Starting new VirtualDatabaseWorkerThread
DEBUG controller.core.Controller Connection from 
e178044177.adsl.alicedsl.de/85.178.44.177:51079
DEBUG controller.core.Controller Starting new VirtualDatabaseWorkerThread
WARN  DatabaseBackend.vdating2.dbnode2 No transaction medatada found for 
transaction 0 releasing locks manually
WARN  DatabaseBackend.vdating2.dbnode1 No transaction medatada found for 
transaction 1 releasing locks manually
DEBUG controller.core.Controller ControllerWorkerThread terminating.
DEBUG controller.core.Controller ControllerWorkerThread terminating.
DEBUG controller.core.Controller ControllerWorkerThread terminating.
WARN  DatabaseBackend.vdating2.dbnode1 No transaction medatada found for 
transaction 2 releasing locks manually
WARN  virtualdatabase.VirtualDatabaseWorkerThread.vdating2 Client 
(login:dating,host:e178044177.adsl.alicedsl.de closed connection with server)

You can see the Connections that have been made bur no errors at all,
only warnings.

There are also some other warnings in it after stating the
Server and foing an "initialize dbnode1 force" and "initialize dbnode2
force" as well as "enable dbnode1" and "enable dbnode2":

WARN  controller.core.Controller Could not parse RMI Server Port 'null
WARN  DatabaseBackend.vdating2.dbnode2 setQueryTimeout test failed.

My Configuration is 2 identical Postgres Databases with RAIDb-1
Loadbalancing.

I noticed that after my first few selects that happen before the
Tomcat is fully operational, i get this line in the JDBC-Log:

"Controller /75.101.243.37:25322 down - shutting down connected
sockets"

Any Request thereafter is answered with Exceptions and the Message
"I/O Error while trying to disable autocommit

It seemd that the Connection is being dropped and not reopened for new
Queries.

Best regards
Manuel Rorarius

------------------------------------
mediaskill OHG
Kaiserdamm 100
14057 Berlin
Fon +49 (0)30 - 39 83 572 - 5
Fax +49 (0)30 - 39 83 572 - 3
  
[EMAIL PROTECTED]
www.mediaskill.de
------------------------------------
Handelsregister: AG Charlottenburg
Nummer: HRA 31638
Ust-IdNr.: DE 238 456 737
------------------------------------

am Freitag, 7. November 2008 um 20:58 schrieben Sie:


> Date: Fri, 7 Nov 2008 11:57:50 -0800
> From: Robert Hodges <[EMAIL PROTECTED]>
> Subject: Re: [Sequoia]  JDBC Connection loss with Sequoia 2.10.10,
>         Hibernate/C3P0, Postgresql 8.3.3
> To: Sequoia general mailing list <[email protected]>
> Message-ID: <[EMAIL PROTECTED]>
> Content-Type: text/plain; charset="iso-8859-1"

> Hi Manuel,

> It looks to me as if Sequoia is taking the backend(s) off-line
> shortly after your application starts.  Have you looked in the log files to 
> see what
> errors appear at this time?  In particular, you should look at
> log/full_cluster.log on all your controllers.  I would expect to see one or 
> more
> error messages that explain why the backends went off-line.

> Cheers, Robert



_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia

Reply via email to