Open GhaziTriki opened 6 months ago
Using vfork
should yield better performance. Is this impetus behind trying out this parameter?
Indeed vfork
can lead to better performance and reduced memory overhead. I am not against the JVM config update but it would require intensive testing to avoid unforeseen issues or conflicts. Maybe 3.0.0 is the right target for a such update.
This is likely caused by the JDK being upgraded (e.g. by Ubuntu's unattended-upgrades
) without the JVM process being restarted. See https://bugs.launchpad.net/ubuntu/+source/openjdk-17/+bug/2055280 for a discussion about a similarly looking problem. https://bugs.openjdk.org/browse/JDK-8325567 also has a fair share of discussion about it.
TL;DR: one should never update the JDK without restarting all JVM processes.
FWIW, vfork
is eventually going away according to https://bugs.openjdk.org/browse/JDK-8214511.
And also, it looks like this will become a permanent feature of the JDK going forward. Any minor update will be considered a breaking change as of https://bugs.openjdk.org/browse/JDK-8325621. It will be part of the upcoming 21.0.4, but has not yet been backported to 17.
Thank you for these updates, @slovdahl !
Describe the bug A Java Exception occured preventing presentations from being converted. The issue disappeared after doing
bbb-conf --restart
Expected behavior There should be no Java Exception.
Possible solution Launch bbb-web with JVM option
-Djdk.lang.Process.launchMechanism=vfork
https://stackoverflow.com/questions/76003646/how-to-fix-java-cannot-run-program-in-directory-error-0-failed
BigBlueButton Server BigBlueButton 2.6.16
Additional context