Hi,
Just trying to understand the Active Agent a bit better.
My results always show that when Hostmonitor is disconnected from the Agent because the link between them is down the test results for the agent show as UNKNOWN. But I know that the test at the agent end should be working.
I though the Active agent was supposed to keep testing even when it was disconnected from the Hostmonitor master host, and then update the master when it came back online. Is this not the case? Have I misunderstood the intent of the Active agent?
I am trying to isolate an intermitent WAN link issue and need testing from both ends to help me narrow down the location and nature of the fault. I was hoping the Active agent would continue to test from the remote end all of the hops so I could quickly identify the offending link/router.
Active Agent - unknown results when disconnected
When connection is dropped, agent is trying to reconnect while HostMonitor does not perform tests assigned to this agent for a while (up to 4 minutes). If connection was not resumed within this time interval while test should be performed due to specified schedule, then HostMonitor sets Unknown status.My results always show that when Hostmonitor is disconnected from the Agent because the link between them is down the test results for the agent show as UNKNOWN. But I know that the test at the agent end should be working.
No, it does not work in such way.I though the Active agent was supposed to keep testing even when it was disconnected from the Hostmonitor master host
Regards
Alex
Actually I still think you should consider such an option. It would be quite handy for the agent to be able to operate in a disconnected mode with a buffer for results. This would be very handy for isolating connectivity issues.
Imagine if the agent cound do a traceroute in the event of a test failure, or just ping test to multiple points in the path. You could then determine where in the link the fault lies and act accordingly. At present all I know is that the agent is unreachable.
There may be other uses too including test that you may want to do on laptops while they are offline
Imagine if the agent cound do a traceroute in the event of a test failure, or just ping test to multiple points in the path. You could then determine where in the link the fault lies and act accordingly. At present all I know is that the agent is unreachable.
There may be other uses too including test that you may want to do on laptops while they are offline
Ok, image we have implemented this option.Imagine if the agent cound do a traceroute in the event of a test failure, or just ping test to multiple points in the path. You could then determine where in the link the fault lies and act accordingly. At present all I know is that the agent is unreachable.
Question: how you will get this information from agent if you cannot reach the agent??
Regards
Alex
Just for log?
I think its does not make much sense. In order to implement this "little option" we need significantly redesign agent. Agent should be able to store test list (just like HostMonitor does), agent should know how to process master-dependant tests, agent somehow should know results of the test items that actually do not belong to agent or belong to different agent (e.g master test can be performed by HostMonitor) and so on and so forth.
At the same time existing "backup rma" option gives you much more advantages than just a log (yes, sometimes primary and backup RMA will be unavailable, e.g. when network does not have backup connection).
Anyway, we rather spend this time for another option - possibility to use several HostMonitors in distributed network so several HostMonitors can work together, balance load and backup each other.
Regards
Alex
I think its does not make much sense. In order to implement this "little option" we need significantly redesign agent. Agent should be able to store test list (just like HostMonitor does), agent should know how to process master-dependant tests, agent somehow should know results of the test items that actually do not belong to agent or belong to different agent (e.g master test can be performed by HostMonitor) and so on and so forth.
At the same time existing "backup rma" option gives you much more advantages than just a log (yes, sometimes primary and backup RMA will be unavailable, e.g. when network does not have backup connection).
Anyway, we rather spend this time for another option - possibility to use several HostMonitors in distributed network so several HostMonitors can work together, balance load and backup each other.
Regards
Alex