Re: james unusable because of delivery failures to postmaster. Configurationproblem?

[Jerome Lacoste @ BBC]

Noel J. Bergman wrote:

I didn't enable special debugging (I can't remember). But there are
plenty of log files in james/apps/james/logs.

In SAR-INF/environment.xml, you can change log levels from INFO to DEBUG.

database-connections.maildb_ ****** connection 22 is way too old: 65063 >

60000

this message being repeated every 60-65 seconds.

java.sql.SQLException: Communication link failure:

java.net.SocketException

at org.gjt.mm.mysql.MysqlIO.sendCommand(Unknown Source) ...

This one is easy to fix, and the fix will be in James v2.1.1.  Change your
config.xml entry to add "?autoReconnect=true", e.g.,

  <dburl>jdbc:mysql://127.0.0.1/mail?autoReconnect=true</dburl>

The problem is that your system is currently overwhelmed in the spool.  I
ran a similar test just yesterday, and saw all of these I/O related errors
before adding the autoReconnect option.

Once you have made that change and rebooted James, it will take a while for
the queue to clear.  You could manually delete entries for the postmaster
account via the mysql client.

thanks for you answers.

I made a DELETE * from spool where recipients = <my postmaster address>;
and it removed most of them.

There is probably another problem as, after having restarted james, I now have 95% CPU usage with mysql.
Perhaps related to those big files in mysql/data/ or to the fact that I remove the contents of the james/var/mail/error directory?

I checked quickly and the bug files in mysql don't seemt to decrease.

Got to go. Will investigate further tomorrow.
Thanks for your help.

J



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]