Sure,

Thanks a lot.
We did sort of performance testing these days: ~7 people came into 1 room
with both audio and video enabled

The server was not very stable on such scenario :( (there were lags in
audio and video, new participants were unable to connect)
I did revert to the previous revision until performance was OK

While this testing we found rev. 1419201 works OK and rev. 1417735 has
performance degradation.
The only major change between these 2 revisions was
ConferenceService.getServerForSession
method, this is why I have added debugging.


increasing of number of publicSID in the cache can be observed by entering
and leaving 1 room by only user
On each entering publicSID count increasing by 3 .... (not sure if it is
expected behavior or not)

Will wait for your changes :)



On Tue, Jan 29, 2013 at 1:40 PM, [email protected] <
[email protected]> wrote:

> Let me check that I have a bunch of changes locally to the SessionStore
> that I have to test before I can commit it.
>
> Sebastian
>
>
> 2013/1/29 Maxim Solodovnik <[email protected]>
>
>> I'm also seeing on server restart .....
>>
>>
>> java.lang.NullPointerException: null
>>         at
>> org.apache.openmeetings.remote.red5.ScopeApplicationAdapter.setUsernameReconnect(ScopeApplicationAdapter.java:1711)
>> ~[openmeetings-2.1.0-SNAPSHOT.jar:na]
>>
>>
>> Maybe you can suggest what is the proper way of handling it?
>>
>>
>>
>> On Tue, Jan 29, 2013 at 1:12 PM, Maxim Solodovnik 
>> <[email protected]>wrote:
>>
>>> Hello Sebastian,
>>>
>>> some time ago I have changed the parent class of ScopeApplicationAdapter
>>> to be MultiThreadedApplicationAdapter (since we should work  multithreaded
>>> ...)
>>>
>>> yesterday I noticed "ApplicationAdapter extends
>>> MultiThreadedApplicationAdapter" so I revert it back.
>>>
>>> Can you please help me with HashMapStore?
>>>
>>> I have added the following code to
>>> ConferenceService.getServerForSession(String SID, long roomId) method:
>>>
>>> log.debug("-- getServerForSession ENTER\n\t\t====================\n" +
>>> sessionManager.getSessionStatistics() + "\n\t\t====================\n");
>>>
>>> as a result I can see
>>>
>>> the following lines in the log:
>>> clientsByServerAndPublicSID Server null Number of PublicSIDs: XX
>>>
>>> where XX is only increasing ....
>>>
>>> Is it by design? should this hashMap be cleaned up?
>>>
>>>
>>>
>>> On Tue, Jan 29, 2013 at 11:47 AM, [email protected] <
>>> [email protected]> wrote:
>>>
>>>> Hi Maxim,
>>>>
>>>> I've seen in your latest changes you have refactored the
>>>> ApplicationAdapter to use the basic interface and no more the
>>>> MultiThreadedApplicationAdapter.
>>>>
>>>> What is the reason behind it?
>>>>
>>>> Thanks!
>>>> Sebastian
>>>>
>>>> --
>>>> Sebastian Wagner
>>>> https://twitter.com/#!/dead_lock
>>>> http://www.webbase-design.de
>>>> http://www.wagner-sebastian.com
>>>> [email protected]
>>>>
>>>
>>>
>>>
>>> --
>>> WBR
>>> Maxim aka solomax
>>>
>>
>>
>>
>> --
>> WBR
>> Maxim aka solomax
>>
>
>
>
> --
> Sebastian Wagner
> https://twitter.com/#!/dead_lock
> http://www.webbase-design.de
> http://www.wagner-sebastian.com
> [email protected]
>



-- 
WBR
Maxim aka solomax

Reply via email to