We are using RedHat 9, Kernel 2.4.20-20.9. JVM 1.4.2_01-b06. We are using Postgres and created our own database connection pooling api. I will try setting the environment variable LD_ASSUME_KERNEL=2.4.1. When you refer to a 'thread dump' are you referring to running netstat and viewing the threads running on the ports Tomcat is set up on? Or is there another way to get a thread dump? (sorry, for my ignorance on linux) Thanks for the help.
-----Original Message----- From: Remy Maucherat [mailto:[EMAIL PROTECTED] Sent: Thursday, December 25, 2003 2:11 AM To: Tomcat Users List Subject: Re: All threads (150) are currently busy, waiting. Increase maxThreads (150) or check the servlet status alan sparago wrote: > I am using Tomcat 5.0.16 standalone serving as both an http/https web server > and servlet container. This error occurred in our QA environment and has > caused Tomcat to hang and not service any requests. This error seemed to > occur when no users were using the site at the time, but about 4 DAYS > earlier we preformed a stress test simulating 1200 semi-simultaneous users > and everything worked perfectly without any errors. Since the stress test we > are only accessing the application as an occasional single user, the last > time being about 8 hours prior to this error happening. Most parameters in > the server.xml file are the default values (the maxThreads value is set to > "150"). > > Our log file is in xml format and contains the following error; > > <logger>org.apache.tomcat.util.threads.ThreadPool</logger> > > <level>SEVERE</level> > > <class>org.apache.tomcat.util.threads.ThreadPool</class> > > <method>logFull</method> > > <thread>13</thread> > > <message>All threads (150) are currently busy, waiting. Increase > maxThreads (150) or check the servlet status</message> > > > > fyi. Nowhere in my application is the code creating any threads. If running Linux 2.4 with tweaks, and esp RH 9, use LD_ASSUME_KERNEL=2.4.1. This kind of report is not very useful without a thread dump once the issue occurs and all threads appear deadlocked, BTW. If there's actually a bug in the Tomcat code, then it's the only way to identify where it is. > We are very close to our site being live so any help or information would be > greatly appreciated. Would tomcat 4.1.29 be more stable? No, it is the same connector anyway. -- xxxxxxxxxxxxxxxxxxxxxxxxxxxxx Rémy Maucherat Senior Developer & Consultant JBoss Group (Europe) SàRL xxxxxxxxxxxxxxxxxxxxxxxxxxxxx --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]