Hi, We're currently running 2.11.5 (upgrade of the whole server scheduled for early next year) and are experiencing stability issues with the Jetty 9.3 hosting Geoserver (and a Geoexplorer). For instance large ortho rasters previewed in Geoserver try to render for a long time and eventually fail and if several of those heavy getMaps happen in parallell Jetty becomes slow to respond and will either end up in state where it doesn't appear to recover, at least for 10 minutes, or it recovers. Lesser rasters preview ok and do not cause
(This is a problem when using these WMS layers in desktop applications, it's not that lots of people preview them from Geoserver's /web) Latest Jetty 9.3, latest Corretto JRE 8. Java options: Marlin 0.9.4.1, -XX:+UseParallelGC -XX:SoftRefLRUPolicyMSPerMB=36000 6144MB initial heap and 8192 max, thread stack size 256KB. When Jetty becomes partially or totally unresponsive the heap via jmapheap -heap looks like so: Heap Configuration: MinHeapFreeRatio = 0 MaxHeapFreeRatio = 100 MaxHeapSize = 8589934592 (8192.0MB) NewSize = 2147483648 (2048.0MB) MaxNewSize = 2863136768 (2730.5MB) OldSize = 4294967296 (4096.0MB) NewRatio = 2 SurvivorRatio = 8 MetaspaceSize = 21807104 (20.796875MB) CompressedClassSpaceSize = 1073741824 (1024.0MB) MaxMetaspaceSize = 17592186044415 MB G1HeapRegionSize = 0 (0.0MB) Heap Usage: PS Young Generation Eden Space: capacity = 1982332928 (1890.5MB) used = 1982332928 (1890.5MB) free = 0 (0.0MB) 100.0% used >From Space: capacity = 437256192 (417.0MB) used = 0 (0.0MB) free = 437256192 (417.0MB) 0.0% used To Space: capacity = 439353344 (419.0MB) used = 0 (0.0MB) free = 439353344 (419.0MB) 0.0% used PS Old Generation capacity = 5726797824 (5461.5MB) used = 5726435656 (5461.154609680176MB) free = 362168 (0.34539031982421875MB) 99.99367590735469% used In addition Geoserver doesn't appear to adhere to the max rendering time under WMS settings: 35s, this is what the Monitor plugin captured shortly before Jetty stopped responding earlier today: /<Request id="3,417,511"> <Operation>GetMap</Operation> <Resources>ortofoto:orto_tatort2018</Resources> <Path>/ows</Path> <QueryString>SERVICE=WMS&VERSION=1.3.0&REQUEST=GetMap&LAYERS=ortofoto:orto_tatort2018&STYLES=&CRS=EPSG:3010&FORMAT=image/jpeg&TRANSPARENT=TRUE&WIDTH=502&HEIGHT=502&BBOX=6575000,141000,6579016,145016</QueryString> <Body> </Body> <StartTime>2019-11-07T07:06:21.469Z</StartTime> <EndTime>2019-11-07T07:13:26.694Z</EndTime> <TotalTime>425,225</TotalTime> <ResponseStatus>200</ResponseStatus> <Failed>false</Failed> </Request> / Geoserver's (production9 or jetty (dev) log do not mention java.lang.outofmemoryerror, but Geoserver will sometimes reply with just that : heap space, when in its unresponsive state and a wfs getFeature is sent. (logging typically stops when the servlet becomes unresponsive). Would a different garbage collector be of use? Setting MaxPermSize? Somehow optimizing the tifs of the mosaic stores? Is there a way to not allow these WMS services to be used except with tiled=true -> their cached variants? -- Sent from: http://osgeo-org.1560.x6.nabble.com/GeoServer-User-f3786390.html _______________________________________________ Geoserver-users mailing list Please make sure you read the following two resources before posting to this list: - Earning your support instead of buying it, but Ian Turton: http://www.ianturton.com/talks/foss4g.html#/ - The GeoServer user list posting guidelines: http://geoserver.org/comm/userlist-guidelines.html If you want to request a feature or an improvement, also see this: https://github.com/geoserver/geoserver/wiki/Successfully-requesting-and-integrating-new-features-and-improvements-in-GeoServer Geoserver-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-users