Scott, *;
first off, thanks a bunch for your hints on that. Much appreciated as I
had almost given up on getting this resolved somehow. :(
Am Thu, 08 May 2008 06:51:25 -0700
schrieb Scott Oaks <Scott.Oaks_at_Sun.COM>:
> The important threads to look at are the httpSSLWorkerThread-8080 --
> those are the threads that handle the HTTP requests. In your case,
> all of those threads are blocked waiting for a response from some
> remote HTTP server. That server is apparently slow, tying up all the
> HTTP threads in your appserver.
Is this a matter of a server being slow or also likely of having large
requests being processed? The glassfish, like most of our environment,
is fronted by a reverse-proxying apache2/mod_rewrite setup, but as we
do document management, most of the activities our clients are involved
in are about transferring large files both up- and downstream. Could
such a "large upload" cause such a behaviour? Should I dump the apache2
front controller and get another proxy setup implemented?
By the way which information in the thread dump makes you consider the
remote server being slow? I'm curious...
> You can add some additional request processing threads for your HTTP
> service, but that's likely only to make things worse; more requests
> to your remote system are likely to make that system even slower.
Yes, that is what I had to learn the hard way already, as well... :(
Anyway, thanks again and best regards,
Kristian
--
Kristian Rink * http://zimmer428.net * http://flickr.com/photos/z428/
jab: kawazu_at_jabber.ccc.de * icq: 48874445 * fon: ++49 176 2447 2771
"One dreaming alone, it will be only a dream; many dreaming together
is the beginning of a new reality." (Hundertwasser)