actually wroot did point my to this direction with his “flood protection” thread.
Spark uses some memory to display images and runs afaik with Xmx64m. I sent some emotions from one client to another and could see how the used memory did increase.
Closing the chat window does not release the memory.
Opening it once again allocates some more memory.
Closing … Opening … Closing … Opening … nearly 100 % CPU usage and if one does this more often (having the time to wait for all the GC’'s to complete) -> OutOfMemory
Any workaround known until this is resolved? We’‘re rolling out ~200 users and we’'re seeing memory usage hitting near 100mb at peak. It does appear to clean itself up from time to time (garbage collection?)… any idea how to trigger this? Perhaps some best practice suggestions we can give to users until resolved?
For a future release of Spark (Spark 2.0), we are making some big changes to the design and frameworks used to lower the memory usage. One of these is moving over to tree instead of our collapsible panes that we use right now. This will allow for only one initialization of a component as well as allow for nested groups.
ex.
– Jive Software
– IM Team
– Forums Team
– Support
– Personal
– Family
– Friends
There are also some other areas of improvement schedule to reduce memory consumption,