skip to main |
skip to sidebar
Compare rates today
Start paying less monthly.
======================================================================================================================== That's funny, I'm using the exact configuration you were using, "Win2K Advanced server with Apache 2.0.40/Tomcat 4.1.10" only with a heavy duty ArcIMS app and I've not seen this. On the contrary, it's been just the opposite for us. The same app under Apache 1.3.26/Tomcat 3.3.1, was eating all the memory. The only real difference was moving from Java 1.3.1 to 1.4.0. rls "craig franke" < crai ... @harlingen.tstc.edu > 09/26/2002 07:54 AM Please respond to "Tomcat Users List" To: < tomc ... @jakarta.apache.org > cc: Subject: Re: Tomcat 4.1.12 memory leak I also experienced this problem using the inline module on a Win2K Advanced server with Apache 2.0.40/Tomcat 4.1.10. Rather than trust the combination I ended up going down to Apache 1.3.26/Tomcat 3.3.1 on a Novell box. For me, setting the -mx paramters did nothing to change the situation. The server still gobbed up memory and resources until it imploded and I had to reboot. Something as simple as holding down the refresh key for 3 minutes was enough to crash it. Not good for a production environment. Craig dami ... @infotechsrl.it 09/26/02 08:41AM >>> Good explanation but it does not solve the problem. Let me try to be clear. - What I am doing is reloading always the same page. - What I see is an ever increasing consumption of memory. Really why should the JVM want more and more ram ? In any case if you keep reloading, this ever increasing consumption eventually will eat all your ram. Whatever is the explanation it does not seem right to me, is there a way to fix it ? Damiano At 09.29 26/09/2002 -0400, you wrote: Each request allocates memory. (And relinquishes accordingly) The garbage collector runs when "it feels like it should". The JVM will continually suck up memory until it reaches its startup parameters. (-mx ...) Once a JVM takes memory from the OS - it does not release it to the OS - it only releases it to its own memory heap. Ing. Damiano Bolla wrote: System: Linux redhat 7.2 Java: /usr/local/j2sdk1.3.1 Tomcat 4.1.12 To reproduce the behaviour you install the 4.1.12 distribution, set the JAVA_HOME run startup.sh and then keep refreshing the homepage http://localhost:8080/ If you monitor the memory usage using top and switching it into display memory usage (Capital M) you sull see tipically something like 22824 22832 22840 23576 23676 23684 23904 23908 23934 23938 ..... This is the SIZE field of the top command. The point is that it never goes down and eventually you run very slowly. Any idea ? Ah, the same behaviour is with jdk 1.4.1 Damiano -- To unsubscribe, e-mail: For additional commands, e-mail: -- To unsubscribe, e-mail: For additional commands, e-mail: -- To unsubscribe, e-mail: For additional commands, e-mail: -- To unsubscribe, e-mail: For additional commands, e-mail: -- To unsubscribe, e-mail: For additional commands, e-mail: I'm curious about the equivalent situation for Wales. I understand we are to have something called Individual Health Records (IHRs) but there seems to be a paucity of information. There's a certain amount of info on the websites of NHS Wales and "Informing Healthcare" but it is largely in marketing speak (ie. aimed at persuading citizens to fall meekly into line) and some of it seems rather out of date. From what I can make out IHRs sound very similar to SCRs but I haven't yet found a satisfactory explanation of the architecture. I understand it has been piloted at the Royal Gwent Hospital, which intrigues me because I recently went there for a minor surgical procedure and there was no mention of IHRs. Can anyone tell me whether there any signifcant differences between the SCR and IHR systems and if so what? It seems they both have the fundamental weakness that they make personal information available via numerous computers in numerous health service buildings where (whatever technical access controls are incorporated) confidentiality is ultimately in the hands of a lot of over-worked, underpaid individuals who are probably sharing login IDs. On 28 Feb 2010 at 15:28, Peter Tomlinson wrote: Not yet everybody. From the article: "As of 19th February 2010, 1,190,418 Summary Care Records had been created and over 6,000,000 patients had been written to as part of a Public Information Programme." Not yet me. Peter Pete Mitchell wrote: I don't recall seeing this at the time ... "Jan 10 2010 Sunday Mail ... Andrew Jamieson sparked a security alert after breaking into confidential computer files on the PM and First Minister and a series of other high-profile Scots ... The breach was discovered on a national database called the Emergency Care Summary system, which holds the details of 2.5million people in Scotland ... there will be no trial as the Crown say that is not in the public interest ... " I'll bet it does. So much for state-of-the-art security. Fortunately Ross is as usual on the ball. See http://www.theregister.co.uk/2010/02/26/nhs_patient_record_privacy_concerns/ Apparently England's summary care record is being rolled out across London and other regions as we speak. Every patient has been notified by post and given a chance to opt out. Except me, of course, but I'm used to that.