NetXMS Support Forum

3 more videos added to the Basics series.

Now taking requests for topics to be covered in the Advanced series!

A few updates today to make the module compatible with 2.0-M releases and forwards.

Changelog:

apt::netxms_repo
- allow specifying which NetXMS apt repo to use

Documentation
- fixed a few minor errors
- updated 'apt::netxms_repo' class documentation


Example, a ping "Local Command" tool:

Normally you would use this as command:
ping %OBJECT_IP_ADDR%

But nothing is stopping you from creating a hard-coded command like this:
ping 1.2.3.4

Or using node's custom attributs like this:
ping %CUSTOM_ATTR_NAME%

You can then set custom attributes on the node as needed... even modify those from within NXSL scripts, so the tool does a different thing under different node statuses... etc...

See this:
https://www.netxms.org/documentation/adminguide/objects.html#object-tools
Look into the "Macro Substitution" section, particularly the last sentance.

You can do all you need to do already on a node object.
Use what Victor said and:

1) Create a NXSL script that calculates the object status as needed.
- NXSL scripts can calculate whatever based on any number of inputs from other nodes, etc...
2) Create a script DCI and deduce node status from the status of this script DCI.
- Victor described this

Tools already support all you need in multiple ways:

1) create a tool that is hard-coded with all details you need.
- filter this tool for the node object you created above
2) Use custom attributes in the tool which would specify what is executed and how
- assign needed attributes to the node object you created above

etc...

Quote from: volkoff on January 24, 2015, 07:41:37 PM
This one is much more serious:
I can't create or delete  DCI.
History:
I tried to play with SNMP's table DCI, but wasn't successful. Later we'll get back to that
But I created one for the node (printer). And since that I can't create or delete DCI. Looks like after "sync" period something breaks, sync thread stops answering to watchdog and I can't connect with console.
So, my opinion - something is wrong in sync thread.
Error is 100% reproducible, so I can supply you with any kind of data.
DB is SQLite, Windows server 2008 R2, but since data is accumulated correctly I do believe something is wrong in sync thread (or with SQLite).

DCI tables do not work with SQLite.

This is a known issue.

As a workaround for now, you can use a different DB.

Quote from: VMGuy on January 19, 2015, 09:37:47 PM
So, today I installed a new 2.0-M1 system and found out this:
...
Tried several times and always the same.

Thanks, VMGuy

This is correct behavior.

As Alex said:

Quote from: Alex KirhenshteinLogin as admin, right click on any object (node / container / object root like "Entire Network" / etc.) in the tree, select properties and set permissions for the user. By default, permissions are propagated to child objects unless you implicitly disable inheritance in object's access control.

So actually give access to the top level objects to the SuperUser group and then it will all work.

UPDATE: Please see this post:
https://www.netxms.org/forum/configuration/interface-names-in-instance-discovery/msg18102/#msg18102

Description: Traffic on {instance-name} - bits/s IN
Parameter: .1.3.6.1.2.1.31.1.1.1.6.{instance}
Origin: SNMP - Unsigned Int 64bit

Transformation: Average delta per sec
Transform script: return $1 * 8;

Instance discovery: SNMP Walk - OIDs
Base SNMP OID: .1.3.6.1.2.1.2.2.1.1
Instance discovery script:
transport = CreateSNMPTransport($node);
if (transport == null)
return false;

// get the interface name to return as instance-name
vb = SNMPGet(transport, ".1.3.6.1.2.1.31.1.1.1.1." . $1);
if (vb == null)
return false;

// filter out loopback
if (vb->value == "lo")
return false;

return %(true, $1, vb->value);

Please note the node has to support the IfXTable for this script to work.

See the "PollCountForStatusChange" server configuration variable.

What version of NetXMS are you using?

Also, SNMP v2c or v3?

3 new videos added to the Basics series today.
Audio should be a bit more leveled between speech and the intro/outro now

Any feedback is always appreciated.

During every configuration poll, if a node doesnt answer on the pre-defined SNMP settings in the node's "Communication" settings tab; all known SNMP communities and user/pass combinations are tested against a node.
If new functional SNMP settings are found, they are applied to the node.

So yes, switch had a bug, but NetXMS behaved as expected.

This is expected from NetXMS and good, since imagine if you needed to change your SNMP community on 100 devices. You would just add your new community in discovery, and NetXMS would automatically find out it needs to use it on all nodes, and you dont have to change the settings of a 100 nodes in NetXMS.
Of course, you would have to reconfigure the SNMP agent on all of the nodes, but NetXMS would pick up the change automatically.

Its a NetXMS feature to use all SNMP communities configured in network discovery.

Here is how it works:
During a configuration poll (which happens when a node is added), ALL v2 communities and all v3 user/pass combinations are tested against a node.
If a working one is found, its used for that node.
This is so you can have different sets of SNMP settings for different nodes - NetXMS will automatically discover what SNMP setting to use for a particular node.

I'm pretty sure its documented in the wiki or the admin manual.

Do you have 'public' as one of your SNMP discovery communities?

Is it possible the node actually responds to SNMP v2 using the public community?

Can you do nxsnmpwalk on the node using v2 public?

Hi,

Thanks for the feedback

• That was not done by me. I got it from Victor for usage in the videos.
• I will fix that in future videos.
• Will pick a better color-scheme if any more work is done in the terminal (I dont think more terminal work will be done in the basics series tho).

More videos coming soon... depending on how these are received, I might do an advanced series.

In the database, go to the "script_library" table and change the script to "return true;".
Server should start now.

Was the script used in any Script DCIs or anywhere else?