NetXMS Support Forum

I have read and watch serveral videos and Im trying to wrap my head around netxms polling methods and how it determines if a node is up or down.  In 1.27 I noticed a status DCI which to my understanding was a basic ICMP check now in 2.0 I dont see that created so what is the process ?  does it depend on if the capabilites of the device ?   If it finds a device that doesnt have an agent or doesnt have snmp  does it continually try these methods when running checks in the future ?  and if a device does have snmp what method does it use to determin status both icmp and snmp or just snmp  .......sorry for long post   

Node status is periodically collected (determined by StatusPollingInterval of Server Configuration).
That status is held internally for each node as one of its parameters.

1) Status DCI was just the node's status collected into a DCI.
You can still create this DCI yourself if you want, or just get current status with a script (right click on node -> "Execute server script").
return $node->status;

2) Status poll checks ICMP, SNMP and nxagent.
You can configure which is used on a per-node basis (right click on node -> Properties -> Polling).

3) If a device uses multiple methods, each method (icmp, snmp, nxagent) is checked.
If SNMP or nxagent do not respond during status poll, proper events (and then alarms according to EPP) are generated.

If you run a manual Status poll on a node, you can see all these checks and their logs in the log of the Status poll.

but if a node has multiple methods which methods generate a node down alarm ? so say icmp works fine but for some reason the snmp agent does not repond   does netxms still consider the node up

Yes, node will be considered up (since it does reply to ICMP after all), and "SNMP Agent is not responding" alarm will be generated.

For example, if a http server is running one the node, and node replies to ICMP as well as HTTP (assuming you have http service monitoring DCIs), it would not be correct to mark the node as down. SNMP service might have crashed, but rest of the services on the node might still be up.

So basically, if node responds to ICMP, its considered UP (but SNMP or nxagent might be down, which will generate corresponding events).
If node does not respond to ICMP, but does to SNMP or nxagent, it is considered UP as well.

If node does not respond to any (ICMP or SNMP or nxagent - assuming all of them are turned on to be used in the Status polling), then its considered DOWN.
(as mentioned previously, you can turn off ICMP or SNMP or nxagent checks during status poll for a node)

got it ... Thank you for taking the time to explain