NetXMS Support Forum

Added a subnet to NetXMS yesterday, ran a discovery on it and then disabled the discovery. This is monitored via a NetXMS agent as a proxy.
It has picked up 45 devices, a mixture of switches, APs and firewalls. I am not really sure exactly how many metrics are being collected in total across the whole network, but I have not specifically started collecting [for example] interface throughput stats or created any custom templates. NetXMS will just be collecting whatever it does by default, things like ICMP reachability, FDB, interface and instance discovery.
In Server Configuration, every parameter with word "poll" or "interval" is at the default except NetworkDiscovery.ActiveDiscovery.Interval and NetworkDiscovery.PassiveDiscovery.Interval which are set to 43000s.

So the reason I am asking this, is that since this network was added, the firewall is showing an additional 2000-4000 connections in to this network from the NetXMS agent. One switch has 64 SNMP connections to it. 19 were opened in the same second. Each connection has <100 TX/RX bytes. So it seems to me that NetXMS is being a bit profligate with how it opens connections.

If I create a new active discovery target and assign a zone to it, the proxy node chosen is "zone proxy" even it doesn't exist. Active discovery will then fail,

Cannot find zone proxy node for address range 10.0.0.0/24@6

but nothing in the UI to indicate this.

"Clear selection" button on "Proxy node" just reverts to "Zone proxy".
I can work around this by using the server itself as the "Proxy".

Version 5.0.3.

I uploaded a bunch of MIBs using the "SNMP MIB files" tool.

The only one that actually works had the extension .mib, all the rest with the extension .MIB are ignored [when you compile the MIBs, it only lists the .mib file].

I guess I can understand that if I just dropped these files somewhere then there may be some rules to follow, but I used the UI to upload these and was given no feedback to indicate that they wouldn't be used. So I think that makes this a bug.

5.0.3, using nxmx-light as a webapp served by tomcat 9.0.58 and openjdk-17-jre 17.0.10+7-1~22.04.1
If I pop out a graph view on a host, I get a server error and stack trace:

org.eclipse.swt.SWTException: Invalid thread access
org.eclipse.swt.SWT.error(SWT.java:3753)
org.eclipse.swt.SWT.error(SWT.java:3676)
org.eclipse.swt.SWT.error(SWT.java:3647)
org.eclipse.swt.widgets.Display.error(Display.java:2309)
org.eclipse.swt.widgets.Display.checkDevice(Display.java:2291)
org.eclipse.swt.widgets.Display.timerExec(Display.java:1127)
org.netxms.nxmc.tools.ViewRefreshController.setInterval(ViewRefreshController.java:92)
org.netxms.nxmc.modules.datacollection.views.HistoricalGraphView.configureGraphFromSettings(HistoricalGraphView.java:382)
org.netxms.nxmc.modules.datacollection.views.HistoricalGraphView.createContent(HistoricalGraphView.java:435)
org.netxms.nxmc.base.views.View.create(View.java:221)
org.netxms.nxmc.base.views.ViewStack.pushViewInternal(ViewStack.java:370)
org.netxms.nxmc.base.views.ViewStack.setView(ViewStack.java:323)
org.netxms.nxmc.base.windows.PopOutViewWindow.createContents(PopOutViewWindow.java:202)
org.eclipse.jface.window.Window.create(Window.java:451)
org.eclipse.jface.window.Window.open(Window.java:834)
org.netxms.nxmc.Startup.createUI(Startup.java:298)
org.eclipse.rap.rwt.internal.lifecycle.RWTLifeCycle.createUI(RWTLifeCycle.java:177)
org.eclipse.rap.rwt.internal.lifecycle.RWTLifeCycle$UIThreadController.run(RWTLifeCycle.java:290)
java.base/java.lang.Thread.run(Thread.java:840)
org.eclipse.rap.rwt.internal.lifecycle.UIThread.run(UIThread.java:107)

Incidentally, Google claims that it has never heard of the NetXMS "Pop out view"!

QuoteNo results found for netxms "pop out view".
Results for netxms pop out view (without quotes):

How are these intervals chosen? Looks a bit random. 1323, 1421, 1520, etc.
Also would be great if there was a marker at 00:00.

Tomcat 9.0.58
NetXMS server 4.4.3
WAR file 4.4.3
Firefox 119 on the client side.

I get various UI glitches when navigating around.

A popup with 'Cannot invoke "org.eclipse.swt.graphics.Color.getRGB()" because "color" is null' will appear at random.

The vertical divider that separates the navigation tree from the main pane of the interface will jump around when clicking in to different icons on the left. It also moves around when resizing the browser window.

There is no scrollbar so at Node level the right-most tabs will simply be off the screen with no way to get to them. Lack of scrollbars applies in other areas as well.

I have a zone that contains:
1x Linux node which is NATed and reached through a port forward on below firewall.
1x firewall on edge
8x switches that use the Linux node as proxy for SNMP/ICMP/etc.

Linux node and switches are being monitored just fine.

I cannot add firewall by public IP because the IP conflicts with the Linux node and duplicate add is prevented.
I can add the firewall with its private IP, but then some hours later, NetXMS deletes it when it realises that the firewall's outside IP is detected as a duplicate of the Linux node.

I could have sworn I'd succeeded at this in the past. I would think it's a fairly common configuration. Do I need to do something "special" to the Linux node to prevent the duplicate detection?

Since updating to 4.4, I no longer see the "Change interface expected state" option. This now has to be done at Properties > Polling, which is only one interface at a time.
Can I expect the old behaviour to come back at some point?

Teltonika SNMP implementation is a bit annoying in that it doesn't populate sysName with the actually configured name of the device. When you install the SNMP package the sysName defaults to 'name'.
I can see that the "Product Name" field detected by the Teltonika driver is actually a useful value, however.
So would it be possible to derive the node name for nodes using the Teltonika driver, from the Product Name?

In address search results, the Switch node is given, if known. It would be nice if I could right-click on the result row and jump straight to the Object Details view of that node. Same for the Node column too, I guess.

In the FDB/ARP cache poll output, a column with the MAC vendor in it would be very useful.

Bit of a big ask this one, probably, but....Vector map exports. Plain SVG would be good but what would be great would be drawio. Great diagram program and it's open-source.

With some Cisco SB switches, if you put them in stack mode then every possible potential interface in a stack is enumerated with SNMP. For example  gi1/0/1 - gi8/0/52 [and the equivalent FastEthernet as well] even if this is a "stack" of just one switch. As an example, I am look at an SG500-28P that has nearly 1000 interfaces, of which only 28 are real physical interfaces, a few are VLANs and the rest are just "imaginary".

They all have an operation state of "notPresent" so should be straightforward enough to filter out, hide or disregard, but can I do this in NetXMS?

Just looking at a server where the NodeDiscoveryPoller queue is over 600000. I can see in the logs that it is trying many many IP address starting with 172.
I found that I had added a discovery target of 172.28.15.0 with a netmask of /0. I don't actually know what this means to NetXMS, was it trying from 0.0.0.0 to 255.255.255.255 inclusive?
I think a warning that a netmask of /0 means way more IP addresses than you'd usually want would be a good idea here.

I find the Object Details > Interfaces view a bit unwieldy due to how many columns there are [I am still scrolling with a 2560x1440 screen], but I appreciate the ability to just export all this info to CSV [useful when assembling a report on a switch].
It would be nice to have a customisable list of fields to choose.
Maybe even some task-focused presets, for example:
- All [obvious]
- Topology: Name, Alias, Description, the 5 Peer* columns, *State columns
- Other use cases I haven't thought of

16:12:21.546 *D* [topology.cdp       ] CDP(sw1 [165174]): remote IP address 192.168.10.5
16:12:21.547 *D* [topology.cdp       ] CDP(sw1 [165174]): node object for remote IP 192.168.10.5 not found

It's found this from the CDP neighbours table, presumably, but there's no hint of this in the UI. It's good information to know about, although could be optional.
Maybe the interface list is not the place for this information, it's already eleventy columns wide and needs another column like I need a hole in the head

Currently have debug level on 5.
Have a /24 subnet with multiple nodes that NetXMS has discovered previously. Two new hosts were added to this network but I just cannot get NetXMS to discover them automatically.
"Active discovery - node 172.22.141.37 responded to ICMP probe via proxy" is logged but that's it, it doesn't log that it took any further action on that specific IP.
If I add the node manually it works instantly, ie, the host is configured with the correct SNMP community.
"Starting SNMP check on range" is logged for the IP range that covers that host, but that's it.
It also mentions that it knows this host from another node's ARP cache.

Cisco CBS series switches are the successor to the Cisco Small Business series switches.
I think it would probably make sense to use CISCO-SB driver for these switches.
Do you want some snmpwalk output or similar to help add support for these switches to the CISCO-SB driver?

If I check "include end nodes" in an L2 map, I get printers and APs.
In my opinion, printers are definitely end nodes and switches are definitely not.
What about APs? I reckon in the context of an L2 map that an AP is not an end node - it provides a L2 connectivity to wireless clients.
How does NetXMS determine what an end node is?

Address Map is a useful tool but I think a few things could enhance it:

• Some way of representing DHCP scopes. I guess this would just be a parameter on the "Subnet" Object class rather than a dummy host that has 50 IP addresses
• Populate it from ARP cache as well
• Colour the above two things with different colours or symbols to distinguish how "true" they might be
• Tooltip when hovering over an IP if it's a known node, click on it to jump to it, etc
• Right click on a "free" IP with option to ping from an agent if there is one in that network with ICMP subagent, and report back if IP was free and no ARP response seen