> I have a 270MB puppetdb-oom.hprof.prev file in /var/log/puppetdb

This isn't unexpected behaviour per se, although it appears as such if
you haven't dealt much with Java applications. Memory usage is a hard
to predict thing, and if its too low, yes the JVM will crash itself &
drop that hprof file. Not all languages have limits, they'll just
increase the usage until the operating system kills them, but Java
does have a hard limit that needs to be set up front, and if it
exceeds that limit we have it set to crash & drop a hprof file in case
we need to analyze it. I don't believe in this case it needs analysis,
not yet anyway :-).

The main bug is in the fact that the daemon stdout isn't being logged
so its easier to determine its happening, at least not for Debian.
Basically Java is expressing the out of memory error via STDOUT, but
STDOUT is not being logged anywhere. We'll have this fixed in a future
release.

For now the recommendation is to increase your heap size until the
behaviour stabilises:

https://docs.puppetlabs.com/puppetdb/2.2/configure.html#configuring-the-java-heap-size

ken.

-- 
You received this message because you are subscribed to the Google Groups 
"Puppet Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/puppet-users/CAE4bNTmzR39EwU5aEhdYLGGq0j2_W0fHy93JFrGo1zdnDyc-cw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to