NetXMS Support Forum

Could you please export or show screenshot of interfaces view? I'm interested only in columns with interface name, description, index, and type. Looks like we have to add physical port mapping on the driver level.

Да, полный старт отмечается сообщением в логе:
2026.06.02 10:44:39.157 *I* [startup            ] NetXMS Server started

Hi!

I've added driver for FortiSwitch devices, it will be part of release 6.2.0 (should be ready in early June). If we will make another patch release for 6.1 we likely backport it too.

Could you please show how hardware inventory view looks like for the switch?
Is "Ports" view already working? If not we might add physical port marking to the driver.

Best regards,
Victor

We just published patch release 3 for version 6.1.

Changes since previous release:
-   Improved AI assistant access to DCI thresholds
-   Improved AI assistant access to alarms
-   AI assistant can initiate polls
-   Improved Qtech driver
-   Improved Cisco SB driver
-   Added driver for Lenovo switches
-   Added driver for BDCOM switches
-   Added driver for Bulat switches
-   Added driver for Origo switches
-   Server no longer attempts to retry agent package deployment after non-transient errors
-   Configurable page size in log viewer
-   Fixed small issues in SNMP printer template

Fixed issues:
-   #3202 (Problems with audit log filtering condition)
-   #3205 (Cropped login form on first run of nxmc on RPi)
-   #3230 (New and changed device drivers)
-   #3232 (Disabled push DCI continues accepting newly pushed data)
-   #3233 (Commands on Overview Windows are shown like gray rectangles)
-   #3234 (.apkg package fails to install under dash due to bash-only [[ ]])
-   #3235 (Display switch IP address in IP/MAC search results)
-   #3236 (Server name is shown in black in GUI header)
-   #3237 (Agent should verify installer signature when upgrading itself on Windows)
-   #3238 (Server crash on shutdown due to use-after-free in the notification channel worker thread)
-   #3239 (Filter values are not set when popping out log viewer)
-   #3240 (Windows agent breaks Store app launches by loading UsrClass.dat of logged-on users)
-   #3245 (AI assistant: add composite 'operational-status' triage tool)
-   #3246 (Threshold line not shown on chart if threshold value is bigger then maxint)
-   #3247 (Thresholds values should be inverted on chart if "invert values" is selected for a data source)
-   #3248 (Filter in Status Map dashboard element is slow)
-   #3250 (Agent tunnel bind fails with error 923 on OpenSSL 3.5.x)
-   #3253 (Empty f-string in NXSL script result in data stack underflow)
-   #3255 (Node without IP address and connected by agent tunnel only never marked as unreachable or down)
-   #3260 (C:\ in file manager incorrectly expanded to C:\Windows\system32)
-   #3267 (Script contents are deleted when dragging "Execute script" view from one docking area to another)
-   #3270 ("Do not show this message again" in "This DCI was added by instance discovery" dialog is not persisted across console restarts)
-   #3272 (No option to configure automatic column resize in DCI summary tables)
-   #3275 (Package deployment jobs should be cancelled if corresponding package is deleted from server)
-   #3276 (Send updates to Package deployment jobs view)
-   #3279 (Cannot upload file to server from nxmc if web server is running on Windows)

Make sure you have --enable-auto-tool-choice set as well. If won't help, try --tool-call-parser=pythonic

I just checked on my system, and topology map definitely takes Topology.DefaultDiscoveryRadius into consideration. Are you sure your map has more than 5 hops from seed node? It is hard to understand from screenshot, too many objects. You can try to set zoom level to 50% or 25% and then do refresh - map will be re-layout to fit screen but at new zoom level.
You can also try to set discovery radius to lower value, like 1 or 2 to check that it is working.

Just closed issue 3023 (former NX-2821) - will be part of release 6.2.0.

That's due to user-configurable tool nature. If all system users use same language, you can just change tool names in server configuration. However, if different language must be supported currently there is no solution. I've registered issue in our tracker for adding localization support to object tools:
https://github.com/netxms/netxms/issues/3243

We just published patch release 2 for version 6.1. Changes since previous patch release:

- Script entry point can be specified in "execute NXSL script" actions
- Improved performance of data migration when upgrading from versions before 6.0
- File manager subagent supports `nofollow` (do not follow symlinks) option for configured roots
- Notification channel drivers that accepts endpoint URLs in configuration or as recipient explicitly forbids any protocols besides http(s)
- Fixed intermittent SNMP proxy failures
- Fixed regression in SNMP walk that broke handling of buggy agents returning table rows in random order
- Fixed pre-authentication buffer overflow in SNMPv3 USM parser
- Fixed bug in updating instance discovery DCIs from templates
- Fixed pre-authentication OOB read in NXCP binary message parser
- Fixed the SHA-1 + AES-192/256 and SHA-224 + AES-256 key extension bug in SNMP library
- Fixed default object tool Connect SSH
- Fixed issues with language switch in nxmc
- Updated Brazilian Portuguese translation
- Added cisco performance monitoring template
- Fixed issues:
  - #2290 / NX-2081 (Allow to go to object details from event log source with right click option)
  - #3195 (Slow query from idata tables on MySQL)
  - #3199 (Add agent metric to read smartctl raw values)
  - #3200 (Port in Ports view should be selected when right-clicking it)
  - #3208 (Add Nodes via ARP Cache Tab)
  - #3211 (Set browser document title for RWT shells)
  - #3212 (filemgr subagent: path-traversal weaknesses in CheckFullPath / GetRealPath)
  - #3214 (Wrong localized texts in Web UI)
  - #3219 (Legacy web API: java.lang.NoClassDefFoundError: jakarta/servlet/ServletContext)
  - #3222 (Dragging of tabs to currenly hidden docking area)
  - #3223 (Allow dragging tabs from main area to docking areas)

Problem is with device processing SNMP GET NEXT requests incorrectly; standard requires that GET NEXT should return OID that is strictly higher than previous one; however, device returns entries in FDB in random order:

1.3.6.1.2.1.17.7.1.2.2.1.2.67.180.5.93.216.211.69 [INTEGER]: 15
1.3.6.1.2.1.17.7.1.2.2.1.2.70.28.131.65.44.65.209 [INTEGER]: 53
1.3.6.1.2.1.17.7.1.2.2.1.2.64.180.5.93.216.211.179 [INTEGER]: 53

third entry actually less than first one.

In version 6 we did some optimizations to SNMP walk, including stricter checks for loop detection, this is why walk on H3C start failing. In next patch release we will relax it to handle buggy devices like this one.

It looks like regression in SNMP library. Could you please try same SNMP walk from command line using nxsnmpwalk tool? Ideally do walk on same device using nxsnmpwalk from v5 as well.
If there will be discrepancies between v5 and v6, send me packet capture from v6 walk if possible.

Is it in desktop or web client?

Regarding AI invocation - I've tried to check qwen3.6-35b-a3b model via OpenRouter, but looks like they don't have providers with tool calling support. From your screenshot it seems that tool call is not recognized and interpreted as part of an answer. Are you sure that runner or service you are using actually follows Open AI API?

Is there really only one row in SNMP walk result? If yes, try to do separate walks on 1.3.6.1.2.1.17.7.1.2.2.1.3 and 1.3.6.1.2.1.17.7.1.2.2.1.2 - there should be more entries. Also try walk on 1.3.6.1.2.1.17.4.3.1.3

Hi,

FDB related code was not changed, it still handles type 6 (secure) in a same way as type 3. Could you please show screenshot of "FDB" page for problematic switch, output of server debug command "show fdb <id>" (replace <id> with actual switch object ID), and result of SNMP walk on 1.3.6.1.2.1.17.7.1.2.2.1?

Best regards,
Victor