Hey,
we have two Xserve, one acting as a failover for the other one. Both
have wotaskd and Monitor running, the first Monitor talking and
configured to talk on wotaskd on the same machine (with the full
QFDN), and the second Monitor is configured to talk to the second
wotaskd. We don't have any instances mixed between servers.
It was working great for almost 4 years, until we got a Cisco
firewall and new switches. Since this move, we get CLOSE_WAIT
happening on some instances. It's never the same instance or the
same application, and the CLOSE_WAIT kill itself after 5-10 minutes
(sometimes longer). We got new Xserve, same problem.
What's happening is that we see CLOSE_WAIT on the Java process, and
the connection was connected to a HTTPd process, but this HTTPd
process don't exist anymore. Sending a kill -QUIT to the instance
don't generate a strack trace, so this is really weird. A kdump
didn't show anything useful, but I was able to see that even if no
requests to the instance are responding, wotaskd can talk to the
instance.
So I was wondering if our problem is related to the fact that we have
two Monitors on the same subnet ? We have two Monitors and two HTTPd
adaptors because since the second server is a (automatic) failover,
it need a complete deployment environment on its own.
What are the symptoms of running two Monitors on the same subnet ? I
don't understand why it would cause problems after 4 years, but I
want to confirm if it can be our problem.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-deploy mailing list ([email protected])
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/webobjects-deploy/archive%40mail-archive.com
This email sent to [EMAIL PROTECTED]
