Thanks for your reply Jeff :-)
I am using it "behind" the Apache Web server (ajp1.3) and am having
serious problems to keep the service alive (not to mention the poor
performance)
The Tomcat is increasing the number of threads when under the load and
inevitable finishes with ThreadPool giving up.
The code in the ThreadPool is really "not okay" as it does not provide any
way of tuning the pools and definetly does not provide any solution when
running out of available Threads :-(
By the way, do you have any experience with the load-balancing under ajp1.3
... up to now I have observed only that one blocked Tomcat is stalling the
whole ajp1.3 load-balancing for the rest of the gang (I have 4 Solaris
machine each configured with identical application).
Also browsing a bit thrue the internet I have found quite a few articles
about Solaris and Threads in JVM ... mostly confusing me if not giving me
very bad feeling in stomach ... any comments about the stability of the
Tomcat3.3/JDK1.3.1_02/Solaris8 ???
Thanks in advance
Drasko
"Jeff Kilbride" <[EMAIL PROTECTED]> on 31.12.2001 06:34:42
Please respond to "Tomcat Users List" <[EMAIL PROTECTED]>
To: "Tomcat Users List" <[EMAIL PROTECTED]>
cc:
Subject: Re: Tomcat 3.3: ThreadPool bug (IS IT SAFE FOR PRODUCTION USE???)
Are you using Tomcat standalone or with Apache? If standalone, then yes, I
would recommend using the latest version of Tomcat 4.x. If you're using
Apache, I would stick with 3.3 until the mod_jk code is fully implemented
with the 4.x versions.
Thanks,
--jeff
----- Original Message -----
From: <[EMAIL PROTECTED]>
To: "Tomcat Users List" <[EMAIL PROTECTED]>
Sent: Saturday, December 29, 2001 2:09 PM
Subject: Tomcat 3.3: ThreadPool bug (IS IT SAFE FOR PRODUCTION USE???)
>
>
> Hi there,
>
> I am trying to deploy our production service on the Tomcat 3.3 platform
> (Solaris8, JDK 1.3.1) but am very often facing the instability problems.
> Somehow, the ThreadPool just runs out of steam and blocks Connectors from
> receiving any more requests.
>
> Looking into the source code, I have noticed some very interesting
> "constructions" in the ThreadPool class. Namely, the setters for the
> maxThreads, maxSpareThreads and minSpareThreads are having very little
> influence as the start() method is resseting the values to the default
> constants ?!?!
>
> public ThreadPool() {
> maxThreads = MAX_THREADS;
> maxSpareThreads = MAX_SPARE_THREADS;
> minSpareThreads = MIN_SPARE_THREADS;
> currentThreadCount = 0;
> currentThreadsBusy = 0;
> stopThePool = false;
> }
>
> public synchronized void start() {
> stopThePool=false;
> currentThreadCount = 0;
> currentThreadsBusy = 0;
> maxThreads = MAX_THREADS;
> maxSpareThreads = MAX_SPARE_THREADS;
> minSpareThreads = MIN_SPARE_THREADS;
>
> adjustLimits();
>
> openThreads(minSpareThreads);
> monitor = new MonitorRunnable(this);
> }
>
> Also, the situation when there are no more available Threads in the pool
is being handled in a special way which JUST BLOCKS IN THE INFINITE LOOP :
> -(((
>
> Is the code of the Tomcat4 equaly immature or there is a fair chance to
> role a productiv system on top of it?
>
> Thanks in advance for any comments
>
> Drasko Kokic
>
>
>
> --
> To unsubscribe: <mailto:[EMAIL PROTECTED]>
> For additional commands: <mailto:[EMAIL PROTECTED]>
> Troubles with the list: <mailto:[EMAIL PROTECTED]>
>
--
To unsubscribe: <mailto:[EMAIL PROTECTED]>
For additional commands: <mailto:[EMAIL PROTECTED]>
Troubles with the list: <mailto:[EMAIL PROTECTED]>
--
To unsubscribe: <mailto:[EMAIL PROTECTED]>
For additional commands: <mailto:[EMAIL PROTECTED]>
Troubles with the list: <mailto:[EMAIL PROTECTED]>
