Hostmon Test Failure
Hostmon Test Failure
I'm running hostmon ver. 4.62 (not as a service) on at NT 4.0 SP 6a, 335 Mhz CPU with 128 RAM. I have 1300+ UNC tests running, where if they fail three consecutive times an external trap sender program is executed and will send a trap to a Unix trap receiver server. This past weekend we had a major power down and shut off most of our servers. Hostmon didn't react too well. Instead of the tests failing, they would just sit at a checking state. I received maybe three failure messages, during a period of 15 hours. Do to this failure, I exported the IP addresses to a text file, wrote a Perl program to use that file, create a .hmi file and import it as ping tests. The ping tests were setup to run once every ten minutes and execute a external Trap sender program when the test failed three consecutive times. This only worked for an hour at a time. After each hour hostmon would report that the ping test failed (which is correct), but would not execute the external Trap program. Restarting the hostmon program would fix the problem for about an hour, then stop excuting the external program again. I think the ping problems might be due to the out dated machine. I normally run the hostmon application on a dual 2.9Ghz PC with 1 Gig ram running XP, but couldn't due to the power down.
HostMonitor's code that performs UNC test does not have any loops. Looks like network client ("Client for Microsoft Networks"?) hangs...Instead of the tests failing, they would just sit at a checking state. I received maybe three failure messages, during a period of 15 hours
Could you check resource usage using standard Task Manager (handles, user objects, threads)? May be there some resource leakage...This only worked for an hour at a time. After each hour hostmon would report that the ping test failed (which is correct), but would not execute the external Trap program. Restarting the hostmon program would fix the problem for about an hour, then stop excuting the external program again.
Actually I would recommend to upgrade that system, I think Windows 2000 is the most releable Windows platform. We tested HostMonitor with Windows NT 4.0 SP3 and SP4. It works fine as well. But system crashed every time we tried to install SP6.
Regards
Alex