Thanks for the advice. I am using a build from the 2.1.x branch so it is a
very recent version. I think the jmap tool will have to be used.. Problem
with that is we now have a system of restarting it so that the OOM doesn't
happen. I guess I need to try to reproduce the issue on dev. But it
doesn't get enough traffic to kill it...
I just wanted to see if there were any big known issues. Sounds like ECW
might be an issue. Some very large extractions are made perhaps they are
responsible.
Next month I should have some time to look at the issue in detail and I
will update you all with what I find.
Thanks again for the information.
Jesse
On Thu, Dec 22, 2011 at 11:23 PM, Rahkonen Jukka
<jukka.rahko...@mmmtike.fi>wrote:
> Hi,
>
> I have seen sometimes with other programs that the ECW driver from ER
> Mapper/Leica reserves whole lot of memory for a while. It made OpenJUMP to
> jam if I had many ECW files open at the same time and I did a lot of
> panning and zooming. The driver is caching some data for speeding up thing
> things if the same area is visited again and it releases the memory after
> 10 or 20 minutes if I remember it right.
> I do not know if this is the same thing. However, all the desktop GIS
> softwares using ECW SDK 3.x based drivers seem to behave in a similar way
> when it comes to memory consumption.
>
> With OpenJUMP it took more memory to use many ecw files as a mosaic than
> when I used only one big ecw file. And JPEG2000 images opened with the same
> ECWJPEG2000 driver behaved in a similar way.
>
> It is possible that this has nothing to do with Geoserver and your problem.
>
> -Jukka Rahkonen-
>
> ________________________________
> Jesse Eichar wrote:
>
>
> > Hi,
>
> > We are having to restart Geoserver once a week because of OutOfMemory
> errors. I wanted to see if anyone knows of particular configurations that
> can cause this. Rough statistics of our server:
>
>
> * 6GM RAM
> * 71 Stores
> * 381 Layers
> * 45 Workspaces
> * Most raster data is ECW and most of those are ImageMosaics
> * 27-30 of the Vector Stores are postgis (We are not yet using JNDI but
> I am trying to find time to set that up). We have set the connections
> perstore to 2 for most of them and have 200 permitted connections so we are
> not running out of connections
> * We have at least one JP2ECW Stores
> * We have a couple proxied WMS Stores
> * Some raster stores are published multiple times in different
> workspaces
> * We heavily use "virtual" workspaces
>
> Any ideas where I should look for the memory leak?
>
> Thanks,
>
> jesse
>
------------------------------------------------------------------------------
Write once. Port to many.
Get the SDK and tools to simplify cross-platform app development. Create
new or port existing apps to sell to consumers worldwide. Explore the
Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
http://p.sf.net/sfu/intel-appdev
_______________________________________________
Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-users