Obtaining a memory dump...

M

Mikhail T.

Hello!

We know, that sending a Java process signal 3 will cause it to dump the
list of threads to the stderr.

What we would like, however, is the (huge) list of objects.

A vendor's program (we don't have sources) runs for a while here
consuming "modest" 80Mb of memory.

Then -- SOMETHING -- happens, and the process' memory consumption starts
to grow rapidly, until it reaches the maximum heap size (currently set
to 3.2Gb) and reports "OutOfMemoryException". The process does not die,
as the exception is caught (and the memory use returns to normal), but
random things begin to fail and it needs to be restarted...

We can not reproduce the problem reliably, so the vendor can't help us
solve it much.

What we'd like is to obtain the dump of all objects to get a hint on
what is causing the out-of-control memory consumption.

I'm aware of multiple different memory profilers, but they all seem to
get the snapshots *upon request*. We can't sit here requesting snapshots
periodically, and doing so will impose unwelcome extra load on the process.

Is there some way, we can obtain a dump, when the exception is thrown?

Alternatively, can we obtain the heap-dump the same way we can obtain
the thread-dump?

Thanks!

-mi
 
L

Lew

Mikhail said:
Hello!

We know, that sending a Java process signal 3 will cause it to dump the
list of threads to the stderr.

What we would like, however, is the (huge) list of objects.

A vendor's program (we don't have sources) runs for a while here
consuming "modest" 80Mb of memory.

Then -- SOMETHING -- happens, and the process' memory consumption starts
to grow rapidly, until it reaches the maximum heap size (currently set
to 3.2Gb) and reports "OutOfMemoryException".

Are you quite certain? I strongly suspect, but have to ask, isn't it actually
an OutOfMemoryError?

The thing is, with an Error you aren't supposed to let the process live:
An Error is a subclass of Throwable that indicates serious problems that a reasonable application should not try to catch.
(<http://java.sun.com/javase/6/docs/api/java/lang/Error.html>)

and yet,
The process does not die, as the exception is caught (and the memory use returns to normal), but
random things begin to fail and it needs to be restarted...

Explaining, perhaps, why the language Designers told the us not to catch that
Error. Tsk, tsk.

Which would still leave you with the difficulty of diagnosing the problem.

Aside from getting a complete JVM dump, is there anything spitting out in the
error messages that gives any clue?
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,755
Messages
2,569,536
Members
45,007
Latest member
obedient dusk

Latest Threads

Top