[IPT] Classloader leak in IPT

Derek Broughton derek at pointerstop.ca
Sat Sep 8 02:21:08 CEST 2012


I have two pure JSP apps (that is, there's absolutely no servlet
code), and GeoServer. The JSP apps have been running for 5 years or
more, Geoserver for at least 3, and I never had a memory issue until I
tried to run IPT

On Thu, Sep 6, 2012 at 10:16 AM, Tim Robertson [GBIF]
<trobertson at gbif.org> wrote:
> Thanks Derek for the info.
>
> What other apps have you got running on tomcat please?  I mean, are you sure this is an IPT issue alone, in which case we will try and replicate your environment and profile?
>
> Cheers,
> Tim
>
> On Sep 6, 2012, at 2:59 PM, Derek Broughton wrote:
>
>> On Thu, Sep 6, 2012 at 4:32 AM, Tim Robertson [GBIF]
>> <trobertson at gbif.org> wrote:
>>> Hi Derek,
>>>
>>> Thanks for taking the time to investigate this.
>>>
>>> Please know that the IPT uses Google Guice which has a known memory leak [ e.g. 1], but is not particularly critical, and we are all waiting for some time for it to be fixed in the next major release.  You might find that this is the cause of the error report.
>>
>> Well, if that's my problem it really _is_ critical, because it brings
>> down my whole Tomcat after a few days.
>>
>>> There have been IPTs known to run for more than a year without issues, so I hope you don't find this to be critical to your deployment - if you find you are doing a lot of redeployments however, then perhaps it would become more problematic.
>>
>> I've had this happening without doing _any_ redeployments. The
>> redeployment just makes it possible to see that the leaks happened.
>> _______________________________________________
>> IPT mailing list
>> IPT at lists.gbif.org
>> http://lists.gbif.org/mailman/listinfo/ipt
>>
>



-- 
derek


More information about the IPT mailing list