Re: [jetty-users] jetty dies periodically

[Simone Bordet <sbordet@xxxxxxxxxxx>]

Hi,

On Fri, Aug 19, 2016 at 5:48 PM, Michael McInness <m.mcinness1@xxxxxxxxx> wrote:
> Apologies for being vague. What I see is thus: I start Jetty with "service
> jetty start". I deploy a WAR (it's a simple CXF REST service), test it,
> debug code, redeploy. Through all of this I'm tailing the *stderrout.log and
> periodically see nothing logged when I redeploy the WAR. When this happens I
> do "ps -ef | grep jetty" and find no JVM running.

Nothing even if you run ps -ef | grep java ? (besides other processes
you may have?)

If so, then looks like the JVM exited.
We don't call System.exit(), so either some other code does, or the JVM crashed.

> Would you classify that as
> a crash or an exit? I'll call it a crexit until I know more.

I guess you're a briton, then ;)

> I do "service jetty start" and it starts back up and the whole cycle repeats. What do I
> expect? I'll take that as a rhetorical question from someone who kindly
> helps others here but grows weary of noobs, as I probably would, too.

It was not rhetorical. Had the process been there, Jetty may just have
stopped working due to a bug you hit, or the heap too full, or the GC
gone crazy, etc.

> I'll enable DEBUG logs and will report back. Thanks for your help.

If it's a JVM crash won't help much.

I recommend that you start Jetty not as a service, but normally from a
terminal and try to reproduce.
If it crashes, you should see that in the terminal. If you don't see a
crash report, it's a normal exit, so next step is a breakpoint in
System.exit() to see who calls it.

Look around for files named "hs_err_pid*.log". They are typically
created when the JVM crashes in the directory you launched the JVM
(not sure what that is when it's started as a service).

What exact JVM version ? Try to be on the latest one (8u102).

-- 
Simone Bordet
----
http://cometd.org
http://webtide.com
Developer advice, training, services and support
from the Jetty & CometD experts.