All,

Just some notes on the progress. May be an interesting read, may be not.
Decide for yourself.

Finally, after a long week or two of debugging, restarting, updating and
tweaking, it seems that the Red5 server is stable and running is a
stress test for a day without crashing, or hanging.

Instead of using indefinite locks, I used a TimedLock that raises an
exception when a deadlock occurs. This saved a lot of hangs, but not
all.

Although it is a bit off topic, it seems there are some undocumented
features in the .Net 2.0 socket handling, which cause random lock-ups of
the whole application... This is really something that drove me crazy.

The server as is is not as CPU friendly as the Java version (yet) since
I go for stability in favor of performance, for the time being.


As soon as this test is concluded, I will get my hands dirty on getting
the latest Red 5 Java and merge the differences into the C# version from
0.62 onward. I hope you guys did not break and change too much :-)

After this, I will try to incorporate spring.net. I use portions of the
java version (converted to C#) that cover what is needed to get it up
and running, but not to my liking.

That's it for now.

_______________________________________________
Red5 mailing list
[email protected]
http://osflash.org/mailman/listinfo/red5_osflash.org

Reply via email to