[fdo-trac] #72: King Oracle v.0.6.5 Memory Leak with MGOS 1.20 Beta2

[FDO]

#72: King Oracle v.0.6.5 Memory Leak with MGOS 1.20 Beta2
-----------------------------------+----------------------------------------
   Reporter:  zspitzer             |       Owner:  haris
       Type:  defect               |      Status:  new  
   Priority:  major                |   Milestone:  3.2.2
  Component:  KingOracle Provider  |     Version:  3.2.2
   Severity:  1                    |    Keywords:       
External_id:                       |  
-----------------------------------+----------------------------------------
 I am using King Oracle v0.6.5 with 10g2 10.2.0.3, looks like it's got
 a memory leak in there somewhere.

 Subsequent testing against mapguide 1.10 show the ram being released but
 the same virtual memory increase is occuring.

 i started to test an oracle spatial layer with MGOS 1.20 Beta2. I am using
 a web/map/layer in session resource. The dataset are the state
 polygons for australia, which contains 12813 records (27 MB in shp
 files).

 Each time I fresh the map, mapguide adds another 30mb ram and 30mb virtual
 mem
 which doesn't get released
 {{{
                                  RAM            VM              CPU Time
 Start Service                   35              31              0.01
 Load Preview                    48              41              0.01
 Reload Map                      123             117             0.27
 reload Map                      155             152             0.52
 reload Map                      188             187             1.18
 reload Map                      221             221             1:44

 }}}

 So back to the memory stuff. I tried the Tasmanian SHP file in
 mapguide which is about 8mb / 5886 rows (complex coastlines and they
 are provided state by state), i kept them as SHP and as usual it's a lot
 faster in SHP land than Oracle and the memory is released after the map is
 rendered. The CPU usage for with SHP nothing compared to the Oracle ( and
 the oracle process was very
 low)

 After about 15 refreshes of the map, memory hits over 600Mb and the
 same in virtual memory. At which point it the mapguide service doesn't
 respond anymore and kill from task manager is needed. there is nothing
 in the mapguide error log either.

-- 
Ticket URL: <http://trac.osgeo.org/fdo/ticket/72>
FDO <http://fdo.osgeo.org/>
Feature Data Objects