tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Carl" <>
Subject Re: Tomcat dies suddenly
Date Thu, 04 Feb 2010 14:58:18 GMT

An interesting situation.  Of the 4028MB of physical memory, top was showing 
usage 4001MB used... now, that would be really inviting.  (Note that none of 
the 19GB swap was being used.)

I had X windows, the system monitor, VisualJVM and a browser running. 
Further, the memory used by the system was constantly increasing (according 
to top) even though Tomcat is essentially the only application (other than 
those I mentioned above) running and, if I understand it correctly, Tomcat 
is (mostly) running in the heap and should, therefore, not be requiring more 
memory (of any consequence.)  I stopped X windows and the applications it 
was running... that recovered about 400MB (according to top.)  However, that 
is slowly being eaten away by something (the pattern is that the used memory 
goes up a bit, flattens and then goes up a bit more... it never goes down 
except for a mment here and there.)  I would have expected the usage to 
flatten out and then fluctuate about that number.  This implies a memory 
leak someplace in the system.

I did the search for 'runtime' as case-insensitive.

If I have a memory leak, cutting the heap in half will just put off the 
problem, won't it?  Also, it would seem the leak has to be outside my 
application as leaks from the application should show up by blowing the 
heap... is this correct?  (permGen is set as 384m, both max and min and is 
currently using around 100m according to VisualJVM.)

Thanks for your insights.


----- Original Message ----- 
From: "Caldarale, Charles R" <>
To: "Tomcat Users List" <>
Sent: Thursday, February 04, 2010 9:14 AM
Subject: RE: Tomcat dies suddenly

> From: Carl []
> Subject: Re: Tomcat dies suddenly
> top and vmstat are reporting 3.6GB...

That must look quite inviting to the OOM killer.

> The only reference to 'RunTime' I could find

It's "Runtime", not "RunTime", if you were doing case-sensitive searches.

> I did cut the Xms, Xmx in half in an attempt to force the problem

That is most likely to eliminate the problem, not force it.  Increasing Xms 
and Xmx should trigger it more frequently.

 - Chuck

MATERIAL and is thus for use only by the intended recipient. If you received 
this in error, please contact the sender and delete the e-mail and its 
attachments from all computers.

To unsubscribe, e-mail:
For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message