- This topic has 24 replies, 6 voices, and was last updated 17 years, 10 months ago by
.
-
Posts
-
Thank you for following up with your workaround for other folks that may be hitting this annoyance as well.
ME folks,
I think you might want to put a small delay (5-10 seconds) into the server restart between the time the server is actually down and the call to start it again. Sometimes the OS needs a bit of time to clean up the IP listening ports the the app server opens up, and attempting to restart a server immediately after it was shutdown can lead to BindExceptions. I’ve seen it more than once, especially on Solaris and Win2K. Both of those OSes had default timeouts for listener ports that were not small.
-jc
JC,
Good point. I was able to reproduce the bind exceptions easily. I filed a bug about it, they do seems to be harmless though. Is this what the rest of you are seeing? Just scary messages in the console, but app server is still running fine?Riyad,
Actually, it’s been my experience that when an app server fails to open a listener, it’s usually hosed. Most don’t seem to have much retry logic in ’em. I’m certainly no Tomcat expert, so I won’t attempt a definitve response, but I’d guess that BindExceptions on startup are not recoverable. And a default Tomcat/JBoss server opens up about a billion ports when it fires up.
-jc
JC,
Could you try this locally? In my tests the odd thing is that I got it to throw BindExceptions right away, and each time I would try and access the different admin web application and other test applications and they worked fine (seemed harmless).Then I handed this off to a developer to fix, and he can’t reproduce BindExceptions to save his life for some reason. I’m trying to figure out the missing link here =/
Riyad,
I restarted JBoss 4.0.4GA on my laptop running Windows XP SP2, and I got a slew of BindExceptions, then tried to access any webapp deployed on it. No luck. I just back blank pages, not even an HTTP 404 or 500. If it helps, it’s a pretty quick machine.
-jc
Another developer on the team was able to reproduce it, it seems to be dependent on how long it takes the app server to shut down, as the re-start may occur before the previous shut-down is complete.
Just a heads up, this will be fixed in 5.0.3. We identified a race condition that can occur during a restart and fixed it.
