New to NetXMS questions - Interfaces and nodes

Started by tomaskir, July 24, 2014, 03:56:23 PM

Previous topic - Next topic

tomaskir

Greetings,

I am migrating our monitoring from Cacti and MikroTik Dude to NetXMS, and I have a few questions. Running NetXMS on Windows Server.

1) In a clean install, the node on which NetXMS server is installed is already added into the "Infrastrucutre services".
If I remove it, it will come back automatically. What is handling this automatic re-adding?

2) The node which is the NetXMS server also has a NetXMS Agent installed on it. Can I disable the agent?
From what I understand so far, NetXMS server does all polling of other nodes through the NetXMS agent on itself anyway. Is this correct?
If so, I should not be able to disable the NetXMS Agent on the server system without losing the ability to poll other nodes. Is this correct?

3) If I add a different Windows node polled through SNMP, and do a configuration poll, it comes up with all Windows interfaces (even the backend Miniports, etc.).
I like this, and I can unmanage these to not have allerts, so that is all good.
But on the NetXMS server Windows machine (which is added by default, and automatically re-added if deleted), even if polled through SNMP, only shows me "real" (with IP address) interfaces it has.
So 2 windows machines, both polled through SNMP show different result, and I do not understand why. How can I see all the adapters for the NetXMS server machine, the same way I see it for a separatelly added windows machine?
Why are these 2 behaving differently?

Thanks in advance!

tom

Victor Kirhenshtein

Hi!

1) It is done automatically by server, usually on startup. If you don't want NetXMS server node under "Infrastructure Services", you can unbind it instead of deleting (right click on infr. services object, select unbind, select server node). That way it will still exist in the system (under Entire Network) and so will not be re-created.

2) Yes, you can. NetXMS server uses own agent for collecting information about itself (interfaces, routing table, ARP cache, etc.) but is able to run without local agent. Local agent only used for polling network services - all standard polls done directly by server.

3) If agent is installed, server prefers interface list provided by agent. Windows agent does not return all those service interfaces as usually it is meaningless.

Best regards,
Victor

tomaskir

#2
Thanks a lot for the help!

1) I was under the impression that all nodes need to be in "Infrastructure Services" - I was understanding the "Infrastructure Services" as a "top level container" of sorts. Would that be correct?
Since I was understanding it like that, I was considering nodes in subnets inside the "Entire Network" only as links to "Infrastructure Services" (since a node will show up in the subnet inside the "Entire network" multiple time)
If so, if I unbind a node in "Infrastructure Services", it will stay as 3 separate nodes inside the "Entire network"?
What about polling, will it be polled 3 time or once?

2) I disabled the NetXMS agent service on the windows server and modified the node in "Infrastructure Services" to poll from SNMP. However, I can still not see the extra interfaces, even if I do a Configuration poll, etc.
How do I get the extra interfaces back after switching from NetXMS Agent to SNMP moniroting? I seems they are somehow ignored.

3) So I can not completely remove the NetXMS server from my NetXMS "inventory"?

4) Coming from Dude, is there a way to setup SNMP profiles? Basicly, all my devices use 2 different SNMP profiles - (one v2c with a comunity, one v3 with auth)
When creating a node, I now have to switch the SNMP version manually and insert the v3 auth secret into each node definition.
Is there a SNMP profile-like functionality, where I can set a bunch of nodes to use a SNMP profile?
What about when I need to change the SNMP v3 secret? Do I have to modify each node separately?

Thanks again!

tomaskir

#3
More questions while implementing more of our services into NetXMS:

Regarding Syslog and SNMP Trap security:
How is security handled in syslog? It seems any host can spam my NetXMS Syslog server. Are there any ACLs for syslog, or do I need to handle this outside of NetXMS?
How is trap receiving secured? It seems that only nodes already added to NetXMS can send traps to the trap collector (with settings for trap taken from the node definition in "Infrastructure services"). Is this the case?

More regarding syslog:
How is the default syslog parsing and assignment to hosts working?
Example: all our routers are properly logging to syslog, and in syslog, source and host are properly paired.
Our HP v1910 switches however, are not getting their syslogs paired properly to their "source" and "host".

Here is the format of the HP switch'es syslog entry:


In NetXMS its not displaying the entire syslog, and not pairing it to the device corectly:


EDIT:
Looking around the forum, the syslog problem is pretty much the same as here:
https://www.netxms.org/forum/general-support/syslog-from-source-unknown/

I can not select syslog format in the HP switch either,

I tried to look in the documentation and searched around for answeres to all these quesions, could not find them :(

Thanks a lot for all the help! :)

tomaskir

#4
Sorry for the question spam, but another:

What are the differences between the different types of Polls (Status, Config, etc...)
Is there any documentation where I can see the differences between the Polls?

I see where to configure timers on each Poll, (Server Configuration), but I still dont know what each poll does :)


EDIT: Nevermind, found this one in the documentation. :)