Url monitors gives false positives
We have an OpManager with a server that has 16 URL monitors configured (bound to a specific server).
We recieve a lot of false positives with the message "Host unavailable" although I have verified using debug-logging on the DNS-server that the actual lookup is successful.
When looking in the log (opmanager_serverOut1.txt), I see a lot of these lines:
19:23:52:256 ServerOut: http://www.host1.com Host Unavailable Connection refused
19:23:52:256 ServerOut: http://www.host1.com Response for url http://www.host1.com is null
or something like this:
12:42:15:071 ServerOut: http://www.host2.com Host Unavailable Premature EOF encountered
12:42:15:071 ServerOut: http://www.host2.com for url http://www.host2.com is Failed to read headers: java.io.EOFException: Premature EOF encountered
It is one specific server that seems to have a lot more problems than any other, which might indicate that it is the actual server that has problem. However, everytime I recieve an alarm and double-check it, the sites never fails. And the actual alarm, "Host unavailable" seems rather incorrect because it is usually a Connection refused. I have also looked for any events in the firewall but don't see anything suspicious.
This is my setup:
Opmanager 8502
64-bit linux, Debian squeeze. Virtual instance on VMWare ESXi 4u1
We recieve a lot of false positives with the message "Host unavailable" although I have verified using debug-logging on the DNS-server that the actual lookup is successful.
When looking in the log (opmanager_serverOut1.txt), I see a lot of these lines:
19:23:52:256 ServerOut: http://www.host1.com Host Unavailable Connection refused
19:23:52:256 ServerOut: http://www.host1.com Response for url http://www.host1.com is null
or something like this:
12:42:15:071 ServerOut: http://www.host2.com Host Unavailable Premature EOF encountered
12:42:15:071 ServerOut: http://www.host2.com for url http://www.host2.com is Failed to read headers: java.io.EOFException: Premature EOF encountered
It is one specific server that seems to have a lot more problems than any other, which might indicate that it is the actual server that has problem. However, everytime I recieve an alarm and double-check it, the sites never fails. And the actual alarm, "Host unavailable" seems rather incorrect because it is usually a Connection refused. I have also looked for any events in the firewall but don't see anything suspicious.
This is my setup:
Opmanager 8502
64-bit linux, Debian squeeze. Virtual instance on VMWare ESXi 4u1