Re: free(): corrupted unsorted chunks
Hi Bernd, Many thanks for your suggestions. It looks like the native APR library was the issue, at least it hasn't crashed yet after removing that library https://tomcat.apache.org/tomcat-8.0-doc/apr.html (the libapr1 package in Ubuntu) and re-configuring the /etc/tomcat8/server.xml to use NIO for SSL, replacing org.apache.coyote.http11.Http11AprProtocol with org.apache.coyote.http11.Http11NioProtocol. I appreciate your help, it saved my bacon! Oliver On 21 June 2018 at 23:34:44, Oliver Kohll (oli...@agilechilli.com) wrote: Aha, I wonder if it's configured with the APR native library for SSL. I will check out that and those things. Thanks Oliver On 21 June 2018 at 21:43:24, Bernd Eckenfels (e...@zusammenkunft.net) wrote: Are you using any native libraries in this VM, is there a hs_err or System core file? Can you run „ldd“ on the java launcher binary. Can you try a OpenJDK binary distribution independent of the OS compiled version? Gruss Bernd -- http://bernd.eckenfels.net
Re: free(): corrupted unsorted chunks
Aha, I wonder if it's configured with the APR native library for SSL. I will check out that and those things. Thanks Oliver On 21 June 2018 at 21:43:24, Bernd Eckenfels (e...@zusammenkunft.net) wrote: Are you using any native libraries in this VM, is there a hs_err or System core file? Can you run „ldd“ on the java launcher binary. Can you try a OpenJDK binary distribution independent of the OS compiled version? Gruss Bernd -- http://bernd.eckenfels.net
Re: free(): corrupted unsorted chunks
e.StandardEngine.startInternal Starting Servlet Engine: Apache Tomcat/8.5.30 (Ubuntu) On 21 June 2018 at 15:31:12, Thomas Stüfe (thomas.stu...@gmail.com) wrote: Whereever you got the "free(): corrupted unsorted chunks" from should contain actually more information. For an example, see: https://www.unix.com/red-hat/233292-free-corrupted-unsorted-chunks.html Thanks, Thomas On Thu, Jun 21, 2018 at 4:27 PM, Oliver Kohll wrote: > Not that I can see I'm afraid - there's nothing else in either > /var/log/tomcat8/catalina.out (or any of the other files in that directory) > or /var/log/syslog. I also can't see anything likely in /tmp > > I'm Googling how to get tomcat to create a stack trace to a file on a crash, > but the consensus seems to be they should be in catalina.out by default. > > On 21 June 2018 at 13:28:46, Thomas Stüfe (thomas.stu...@gmail.com) wrote: > > Hi Oliver, > > Is there a stack trace on stderr? > > Thanks, Thomas > > > > > > On Thu, Jun 21, 2018 at 1:59 PM, Oliver Kohll > wrote: >> Hi, >> >> I'm a Java developer getting a crash and error message 'free(): corrupted >> unsorted chunks' in the log. this sounds like it could be a low level >> issue >> in the core Java libraries so I'm posting here, but if it would be better >> in another list (or even a different community) please let me know. >> >> So I'm requesting help because I don't know how to debug this further. In >> summary, the issue is >> >> We're getting a crash nearly once a day on our production system, at a >> time >> of day when it's most busy (mid morning). This crash hasn't been seen on >> our development environment, which is the same apart from the server has >> less memory and CPU resources. >> >> The environment is: >> * A VPS host >> * Ubuntu LTS 18 (Bionic) >> * The default Java 10 packaged as 11 I believe ( >> >> https://askubuntu.com/questions/1037646/why-is-openjdk-10-packaged-as-openjdk-11 >> ) >> * However, javac seems to be from JDK8: >> root@server:~# file /etc/alternatives/java /etc/alternatives/javac >> /etc/alternatives/java: symbolic link to >> /usr/lib/jvm/java-11-openjdk-amd64/bin/java >> /etc/alternatives/javac: symbolic link to >> /usr/lib/jvm/java-8-openjdk-amd64/bin/javac >> >> We're running our webapp on Tomcat 8 (also the Ubuntu default). When this >> happens, the server stops responding to HTTP requests (in fact the java >> process has quit) and needs to be restarted. >> >> The crash doesn't seem to correspond to a particular activity on the >> server >> as far as I can see, though it's hard to tell as many people are logged in >> at the same time. >> >> Does anyone have any pointers of >> a) how to track the bug further so I can report it, ideally without a big >> impact on our live server >> b) more immediately, any workarounds to try >> >> In the meantime I will see if I can downgrade to JDK8 on the test server, >> but I'm not sure if that's possible or whether it would make a difference. >> >> Regards >> Oliver
free(): corrupted unsorted chunks
Hi, I'm a Java developer getting a crash and error message 'free(): corrupted unsorted chunks' in the log. this sounds like it could be a low level issue in the core Java libraries so I'm posting here, but if it would be better in another list (or even a different community) please let me know. So I'm requesting help because I don't know how to debug this further. In summary, the issue is We're getting a crash nearly once a day on our production system, at a time of day when it's most busy (mid morning). This crash hasn't been seen on our development environment, which is the same apart from the server has less memory and CPU resources. The environment is: * A VPS host * Ubuntu LTS 18 (Bionic) * The default Java 10 packaged as 11 I believe ( https://askubuntu.com/questions/1037646/why-is-openjdk-10-packaged-as-openjdk-11 ) * However, javac seems to be from JDK8: root@server:~# file /etc/alternatives/java /etc/alternatives/javac /etc/alternatives/java: symbolic link to /usr/lib/jvm/java-11-openjdk-amd64/bin/java /etc/alternatives/javac: symbolic link to /usr/lib/jvm/java-8-openjdk-amd64/bin/javac We're running our webapp on Tomcat 8 (also the Ubuntu default). When this happens, the server stops responding to HTTP requests (in fact the java process has quit) and needs to be restarted. The crash doesn't seem to correspond to a particular activity on the server as far as I can see, though it's hard to tell as many people are logged in at the same time. Does anyone have any pointers of a) how to track the bug further so I can report it, ideally without a big impact on our live server b) more immediately, any workarounds to try In the meantime I will see if I can downgrade to JDK8 on the test server, but I'm not sure if that's possible or whether it would make a difference. Regards Oliver
