Hi Mark
Whenever you see JVM hang you need to follow 2 things: check garbage
collector log to see if there's anything bad going on (alternatively use
jstat -gc PID 1000 5 which will print 5 lines of cumulative gc statistics
in 1 second intervals, and if GCT is constantly increasing, you have a gc
problem), and get a few thread dumps with jstack PID. You can use a script
I wrote some time ago to aggregate thread dumps which may help to see
the issue
https://www.pythian.com/blog/a-simple-way-to-monitor-java-in-linux/
If you prefer gui then try jvisualvm which will show you gc activity
visually, and other stuff as well.
On Friday, 30 September 2016, Mark J. Bobak <mark@xxxxxxxxx> wrote:
Hi all,
Our standard stack is Oracle Database 12.1.0.2->APEX 5.0->ORDS 3.0.6->
GlassFish 4.1.1.
This generally works well, and is implemented on several sites (possibly
with minor version variations on various components.)
However, on one particular server, Glassgish periodically slow down, and
eventually hang to the point that, when trying to reach the app login
screen in a Web browser, we'll get connection reset errors.
I *thought* this was the JDBC & /dev/random entropy pool problem. But
yesterday, I modified GlassFish JVM configuration to use /dev/urandom, and
users continue to report intermittent slowness and connection reset errors.
I'm about at my wits end here.
Help!
If anyone has any ideas/thoughts/suggestions, I'd love to hear them!
AdvThanksance,
-Mark
