[mapguide-users] Here's some stress testing on Mapguide

[Bruno Scott]

I have found one memory leaks, but it was my fault
in MapFrame.aspx i was testing if the session://.../sheboygan.Mapdefinition
was existing
            try
            {
              MgByteReader tmpReader =
resourceSrvc.GetResourceContent(mapStateId);
            }
            catch(MgException mge)
            {
              mapStateIdExist = false;
            }
In my test case the session map resource was never created before so it went
to the catch wihtout disposing anything, this was causing a memory leaks on
the w3w service. I was using this because the MgMap.Exist function was not
working in release 2.02. Maybe it is working now, i will check later.
So wihout this little line of code we got these results with 20 concurent
users looping on these 3 test
   http://.../mapagent/mapagent.fcgi?OPERATION=CREATESESSION...
   http://.../mapviewernet/mapframe.aspx?MAPDEFINITION=Library://...
   http://.../mapagent/mapagent.fcgi?OPERATION=GETDYNAMICMAPOVERLAYIMAGE&...

http://n2.nabble.com/file/n3200253/Image1.gif 

we see that Mapguide memory is stable after 5 minutes and the w3w service is
quite stable.
With this test we can't conclude on any memory leaks

Now back to the the full test, with loading ajax viewer,select, zoom, pan
Same server, same sheboygan map and still 20 users

http://n2.nabble.com/file/n3200253/Image2.gif 

Here it is quite clear that we still have a pretty big memory leaks on both
Mapguide and IIS
My own dotnet error on the mapframe.aspx gives me a place to look for.
I don't see lots of Dispose in the aspx ajaxviewer pages.
I know that DotNet have a garbage collector that suppose to get ride of
these memory problems
It does seems to work fine ( to be polite... )

Still investigating
Bruno
-- 
View this message in context: http://n2.nabble.com/Here%27s-some-stress-testing-on-Mapguide-tp3182666p3200253.html
Sent from the MapGuide Users mailing list archive at Nabble.com.