gump-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefan Bodewig <bode...@apache.org>
Subject Re: Disabled Nagging on vmgump
Date Tue, 17 Aug 2010 19:33:21 GMT
On 2010-08-17, Stefan Bodewig wrote:

> On 2010-08-17, Stefan Bodewig wrote:

>> Instead of being helpful it seems to now terminate all forked VMs
>> after a certain amount of time - unless anything else is responsible
>> for this, that is.

> Linux's oom-killer is repeatedly terminating java and python processes.
> Looking through the logs I see it has happened in the past as well but
> happens way more often this time around.

After the Gump run was finished there still where about ten or fifteen
java processes active, all of them running
surefire-booter-something.jar.  They may be leftovers from when Gump
terminates a build process, I'll keep an eye on it.

After I killed the processes the machine immediately reclaimed several
hundred MB of swap space.

The current run has been able to build ooxml-schemas, one of our bigger
memory consumers and memory is no real issue (still about 1.8GB of swap
available).

Given the machine has 1GB of physical memory and 2GB swap space we
should be able to finish Gump withou any memory upgrade.

I'll revisit the surefire-booter case and add a cron job to kill such
processes if necessary.

Stefan

---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@gump.apache.org
For additional commands, e-mail: general-help@gump.apache.org


Mime
View raw message