NetXMS Forum

The device(s) that are being detected with that 10.10.10.56/29 should be showing up under that subnet folder so you should be able to narrow down the culprit...

Check this thread:

https://www.netxms.org/forum/general-support/script-execution-error/

Are you able to poll other SNMP nodes through the same agent/proxy?

Have you done a "tcpdump port 161" on the agent while attempting the poll to ensure there is expected traffic?

If you're running selinux or something equivalent on the agent/proxy?  You might want to disable it and try again to see if this is the cause.

-Zeb

Your script works without issue on my end.  I'm using 3.8.314.

alarm = FindAlarmByKeyRegex("SERVICE_DOWN_.*");

works for me, but only if there are alarms that have SERVICE_DOWN_.  If there aren't any, you'll get a "Function or operation argument is not an object" if you try to println alarm->message because alarm isn't an object.  Alarm would be set to NULL in this case, so best to check for this:

if (alarm != NULL)
  println alarm->message;

We had a similar challenge with the poller queue never "catching up" and would continue to grow.

Victor took a closer look.  Turning on parallel processing and increasing the Discovery.BaseSize from 1 to 8 seems to have helped immensely.  We haven't had any related problems since.

Not sure if this will help your situation.

-zeb

Hi,

Fairly new to NetXMS and running the latest version 3.8.193.  Have been running it since 3.8-ish and am just now digging into some of the layer 2 topology mapping.

We use Mikrotik switches almost exclusively and I've noticed that NetXMS can't seem to find the port a device is connected to on a Mikrotik switch.

For instance, at home I have a simple network configuration:  Mikrotik CCR Router -> ProCurve Switch -> Mikrotik CRS324 switch -> NetXMS Agent PC.

If I right-click on the Agent PC and select "Find switch port", it tells me it is "indirectly connected to Procurve Port 5".

If I right-click on the CRS324 switch and do the same.  It also tells me it is "indirectly connected to Procurve Port 5".

The topology maps also tend to be a bit goofy and I suspect whatever is causing the above, is also causing the mapping issue.

Any idea where I should start looking?  I'm running RouterOS 6.48.1 on all Mikrotik devices, and they all have a bridge configured for switching (with hardware offloading).

Thanks!

-zeb

We have the same issue, but we use a lot of Zones with Proxy Nodes and a lot of overlapping subnets so using the "Accept node if it's within given range" can be cumbersome to manage.  The only way I found so far to prevent scanning in our case is to add firewall rules on the edge device that prevents connectivity and scanning to that subnet, from the proxy node.

If there's a better way to do this, I'm all ears!

Thanks Filipp!

If you need any additional details or a guinea pig for testing, just let me know.

We're really starting to get deep into NetXMS and the more we use it, the more we're amazed.  Great work.

3.8.166 still seems to have the same issue.  Should I open an issue on github?

I just got this working yesterday, at least through a script, and found this NXSL command helpful:

SendNotification("Notification Channel", "Channel", "text", "text2");

In your case, this would be:

SendNotification("Teams Notification", "NetXMS", "does this work?", "Yup!");

I ran it manually through "Execute server script".  Worked great.

I haven't tried it as an actual event notification yet.

-zeb

If the system you're doing the ping on is something like Linux, then "ping" never actually stops to give output with it's default commandline.  You'll want whatever you run through the SSH subagent to return a final value/string that you can deal with.

For instance, on Linux, a better ping command would be something like:

ping -qc1 192.168.240.1

And then parse it with awk or grep to give you a single line output.  There are some good ideas in this thread:

-zeb

You have the same issue.  You're running Java 8.  Install Java 15 from the link above (or java 11 if you prefer) and then UNINSTALL Java 8.  Then you should be good to go.

NetXMS uses the eclipse framework for Java.  Newer versions of the framework require Java 11 or higher, which is why you're getting the error.

When you get the error again, check the "java.version" near the top of the log.  It should indicate java 11 or higher.  Your original log shows java 1.8.0 144 (ie. JAVA 8).

-zeb

Thanks Filipp,

What's the best practice for updates when using agents with deb packages then?  We can't use the Package Manager in NetXMS for these updates and will need to rely on the OS update methods?

I had to remove java 8 for java 15 to be the default.  This was on windows.

Just upgraded to 3.8.120 and still the same issue.

The only workarounds so far are:

• Enable SNMP on the same machine that's running the agent.  It will detect the node, add it, and then also discover that it has a NetXMS agent installed.

• Turn on "Active Only" network discovery.
• Turn off all filters.
• Add ONLY the IP(s) of the agent(s) you want to add, and do a scan.
• Once added, remove the IP's from the active list.
• Turn SNMP and NetXMS Agent filter back on.
• Set network discovery back to "Passive Only".

-zeb

@filipp-sudanov said in Web Service DCI howto?:

netsvc should not be required for web services. It should be present in netxms-agent package, please check output of
dpkg-query -L netxms-agent
If its not there, what is the version of OS?

No, nxget, nxaction, nxwsget do not support tunnel connection - they attempt to connect to agent directly.

Hi Filipp,

We're running mostly on CentOS so we use the tar/APKG packages and there isn't a "netxms-agent" package installed per se.

If I do a nxwsget from the server to one of these agents, I get a "501: Not implemented" error.

I'm not sure what I'm missing.

I did get web services to work on this agent though, just not using nxwsget and the example you provided (nxwsget <ip of agent> http://api.open-notify.org/astros.json number)

-zeb

@victor said in Web Service DCI howto?:

@zebble said in Web Service DCI howto?:

Finally, slightly related but is the asterisk.nsm subagent available anymore?  It appears that the libosip2/libXosip2 library dependency is really old and unsupported with newer version of openssl, so is difficult to build on more modern build environments.

Asterisk subagent is supported. We provide binary builds for all supported Debian and Ubuntu versions. Ubuntu 20 does not have libexosip2 package, but when you install Asterisk subagent from our repository it will bring our build of libexosip2. If you build from sources, you can download and build libosip2 and libexosip2 version 5.2.0 - they are fairly new and seems to be building without issues on modern distributions.

Best regards,
Victor

For the most part, we're using CentOS so we use the linux x86_64 tar for installation, and the APKG packages and the Package Manager for subsequent updates.  The APKG doesn't appear to have the SSH or Asterisk (or some other subagents) installed.  Any chance these will get added to the APKG and tar installation/update methods?

I was able to build the Asterisk subagent finally after searching for libosip2 and libexosip2 version 5.2.0 as you noted.  Not very easy to find!  The subagent works really, really well so far.  Gives us a lot of metrics that often require very expensive tools to monitor.  Nice work!

-zeb

I agree with that.  I seem to remember Java 8 working for early versions of 3.7 and then stopped working for newer versions.  Not sure at which version this changed, nor did I see any notes on the download page that Java 11+ was required, so it was a bit confusing.