Description of problem: Attempting to talk to a netdump server on a non critical port, creates the error messages "Got too many timeouts in handshaking, ignoring client". Version-Release number of selected component (if applicable): netdump-0.6.11-3-i386 netdump-server-0.6.11-3-i386 How reproducible: Steps to Reproduce: 1. Setup netdump to dumb on non standard port 2. Test basic netdump logging 3. Panic machine Actual results: Logs hit disk, core never hits disk Expected results: Core to hit disk. Additional info: Workaround to use a standard port, but this is not available/a solution in this customers situation. Customer has been informed that this does not meet inclusion criteria for RHEL3U9 as this is clearly in maintenance mode.
Yeah, sorry, this is a wontfix, although theres not enough data here to really analyze the problem. Best suggestion would be to go get the netdump-server package from EPEL for RHEL5 and retest with that. If that fails in the same way, open a bug on that, or reopen this one against that component, and we can track down the problem. FWIW, I vaguely recall a bug that we fixed in RHEL4 for netdump regarding non-standard port usage for netdump, but I think it required that the crashing machine was also acting as a netdump server