EIF probe hanging

Hi List,

We are having a growing problem with our EIF probe hanging. This has been occurring about 5-10 times per day. We are running on a Linux RH5 VM, 2 CPU, 4GB mem. Object server is on the same box, loadavg is typically about 0.50. Normally we get about 110k events/day, and usually have about 65 sustained connections (mostly mainframe 'event pump'). When we experience a probe hang, it is as if the comm on the EIF port freezes. We start seeing connections on the probe server going into the SYN_RECV state, and if we attempt to connect to the probe port from another box the comm hangs as if there is a firewall, and you'd see a SYN_SENT state on the connection attempt. Connections to other ports on the probe box work fine, and connections to the probe port work fine AFTER we bounce the probe.

I would not have thought that an application listening on a TCP port could (by itself) affect comm to this degree, so that a tcp handshake can't even take place, so I am suspicious of a network driver or configuration issue on this server.

Has anyone seen any issues like this with the EIF probe? Any thoughts on narrowing the problem?

Version info (from top of eif log):

Netcool/OMNIbus Probe API Library Version 7.3.1
API Release ID: 5.21.27
Release ID: 10.0.10

TIA,

-James

_______________________________________________
TME10 mailing list
[email protected]
Unsubscribe:[email protected]