I have recently investigated a problem with our Application Centre setup (
2
machine nodes) where some COM+ components were:
1) not reusing ****ts, so each request would seem to grab another;
2) not being "idled" so would stay alive indefinitely.
The node would then eventually run out of IP ****ts.
I discovered "Remote Server Name" was blank (on both nodes). After setting
this to the cluster name on both nodes the problem seems to be resolved.
Why was this happening, and in particular, why only on one node? The
component on the "good" node seemed to be behaving correctly?
thanks for any reply
bruce
