I put the test build to run in our server and so far the problem did not happen again
the connection log shows a smaller number of connections than netstat what is correct
Usually I would aready have more connections in the connection log than in netstat after
12 hours of server up.
I will keep check until tomorrow, if everything is fine I will install this version in
all of our server
[]s
On Thu, 2004-04-15 at 00:57, Noel J. Bergman wrote:
I encountered this problem tonight, and have opened up a JIRA issue to track this problem: http://nagoya.apache.org/jira/browse/JAMES-254 Unfortunately, I reset the server before checking the state. I did review the logs, and I have a few ideas about what might be causing it, but there wasn't sufficient information in the logs to be conclusive. I am making some changes to the logging, so that exceptions that terminate the connection are logged at ERROR instead of DEBUG level. Some of you have apparently been encountering this problem all-too-frequently, and already have DEBUG enabled for the SMTPHandler. What I'd like for you to do is check the smtpserver log to see if you have a bunch of messages of the form: $DATE ERROR smtpserver: SMTP Connection has idled out. but without a subsequent exception, which might be of the form: $DATE DEBUG smtpserver: Socket to $remoteHost ($remoteIP) timeout. Also, if the connection log says that you are maxed, and netstat --inet does not show that many connections, do a KILL -3 on the james process so that we can get a thread dump, before you restart James. Right now I am letting a connection idle out before the HELO to see what it does. If that works properly, I'll let it idle out after the HELO, and will walk my way through until I reproduce the problem. --- Noel --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
