It does seem to have something to do with the MemTracker, so that likely
would fix it. Could you let me know if it makes it into a 4.5 release?
--
Jordan Walker
Center for Integrated Data Analytics
US Geological Survey
8505 Research Way
Middleton, WI 53562
608.821.3842
http://cida.usgs.gov
On Wed, Feb 25, 2015 at 6:10 PM, Christian Ward-Garrison <cwardgar@xxxxxxxx>
wrote:
> Hi Jordan,
>
> I haven't investigated this specific issue, but I suspect it's related to
> this: https://github.com/Unidata/thredds/issues/97
>
> If so, it's been fixed in the development branch, 4.6.0. We may apply it
> to 4.5 as well.
>
> Cheers,
> Christian
>
> On Wed, Feb 25, 2015 at 3:52 PM, Walker, Jordan <jiwalker@xxxxxxxx> wrote:
>
>> We are seeing something on TDS 4.5 that hadn't been occurring on 4.3.
>> Our heap is set to 4GB, but with somewhat heavy use this is filling up very
>> fast (several hours) and causing out of memory errors. Looking briefly at
>> some heapdumps. It appears that the issue is some hdf5 headers that are
>> sticking around in the heap and not being cleaned up. Specifically there
>> are about 5,500,000 char[] that are around 500 bytes each.
>>
>> Is it possible that I have a misconfiguration that is causing this to
>> occur? Otherwise, what else might I provide that will help diagnose this
>> problem?
>>
>> --
>> Jordan Walker
>> Center for Integrated Data Analytics
>> US Geological Survey
>> 8505 Research Way
>> Middleton, WI 53562
>> 608.821.3842
>> http://cida.usgs.gov
>>
>> _______________________________________________
>> thredds mailing list
>> thredds@xxxxxxxxxxxxxxxx
>> For list information or to unsubscribe, visit:
>> http://www.unidata.ucar.edu/mailing_lists/
>>
>
>
