[omniORB] Exception trying to start new thread

[Duncan Grisby]

On Tuesday 4 April, radamkie at kdm.pl wrote:

> By optimal settings of stack size and parameters in omniORB.cfg, I
> made test with max load of requests between applications to reach
> limit of threads.  There wasn't "Exception trying to start new
> thread", but the client application reach 100% of processor
> utilization and the serwer application after some time aborted with
> such message in trace:
> 
> POAPOAPOAPOAPOAPOAPOAPOAAborted

That looks like multiple threads all tried to log something, then it
died. I suspect you've run out of stack, which caused it to all go
wrong.

> Also, when I stoped the client application, by terminating, I receive
> many such messages:
> 
> omniORB: Disable ObjRef(IDL:org/csapi/cs/IpAppChargingSession:1.0) root<3>
> omniORB: omniRemoteIdentity deleted.
[...]
> omniORB: Disable ObjRef(IDL:org/csapi/cs/IpAppChargingSession:1.0) root<3>
> omniORB: omniRemoteIdentity deleted.
> 
> Why there is so many omniRemoteIdentity to delete? Because of too many
> threads?

Nothing to do with threads. Each object reference has an identity object
associated with it. Those log messages are just telling you that the
identities are being deleted as the object references are disabled
during shutdown. The messages show that you have lots of object
references present when you shut down. If you expect to have lots,
that's fine. If you think you shouldn't have them, you have a reference
count leak causing them to stay around when they should be gone.

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --