[omniORB] servant refuses serving, "Exception trying to start new thread (8)"

[Igor Lautar]

Hi,

Using omniORB 4.1.1 with VS2005.

We have encountered a problem where the following occurs (omniORB
debugs):

Server accepted connection from giop:tcp:<ip>:3021
Exception trying to start new thread (8).
Cannot create a worker for this endpoint: giop:tcp:<ip>:23231 from
giop:tcp:<ip>:3021

(ip is the same for client and servant).

Client then receives COMM_FAILURE.

I've only seen this once and suspect that it has something to do with
resource
leak (probably in our own code), servant in question has been running
for a month.

After quick examination, error 8 seams to come from GetLastError() on
windows,
being not enough memory, which would suggest memory leak (again, most
probably in our code).

Is this the likelly reason or is there some other possibility related to
omniORB itself?

Regards,
Igor