On 12/05/2012 04:07 PM, [email protected] wrote: >> Are you running the gateway virtualized or on real hardware? > > This is real Hardware and no memory shortage either. Djigzo was working > as expected at the time it was killed. Today where it was noticed i was > even working in the admin console, so it should be no stall of Djigzo > but more of a false restart trigger.
Strange. I have never seen this problem on my own servers (well at least not that I know, since it automatically restarts :) I have checked the Java Wrapper releases notes at: http://wrapper.tanukisoftware.com/doc/english/release-notes.html Although 3.4.1 is already old, I did not immediately see a bug fix which might explain what happens. This looks the closest: Version 3.5.9: Fix a problem where the Wrapper was not correctly yielding control back to its main loop when very large amounts of continuous output was being logged from the JVM. Introduced in 3.4.0. In versions prior to 3.5.8, this could have caused the JVM to timeout and restart itself. That particular issue was resolved but the Wrapper process in 3.5.8 would still have been unresponsive when this was happening. The Wrapper will now always yeild back to its main loop after 250 milliseconds of continuous logging. You might try to use a newer Java wrapper release to see whether this fixes the problem. The Java wrapper files are stored in: /usr/share/djigzo/wrapper The wrapper tar is extracted to a sub dir and then the following symlinks link to the required files inside the version specific sub dir: libwrapper.so --links-to--> wrapper-linux-xx/lib/libwrapper.so wrapper-djigzo --links-to--> wrapper-linux-xx/bin/wrapper wrapper.jar --links-to--> wrapper-linux-xx/lib/wrapper.jar If you want to test with a newer release of Java wrapper, untar the file and overwrite the symlinks and then restart djigzo. Kind regards, Martijn -- DJIGZO email encryption _______________________________________________ Users mailing list [email protected] http://lists.djigzo.com/lists/listinfo/users
