NetXMS Support Forum

We just published patch release 7 for version 5.2. It fixes some important bugs, along with some improvements. Full change log since previous patch release:

- Subagent for monitoring Redis database
- Optimizations in server process memory usage
- Fixed bug that cause completed system scheduled tasks not being deleted from database
- Fixed incorrect behavior of NXSL functions Math::Min and Math::Max when first argument is an array
- Fixed server crash on zero period query for business service uptime
- Fixed incorrect VLAN list retrieval on some Juniper switches
- Fixed issues:
  - NX-2594 (Possible server deadlock in network discovery)
  - NX-2652 (Improve overview page for interfaces)
  - NX-2813 (duplicate key value violates unique constraint "interface\_vlan\_list\_pkey")
  - NX-2836 (Unbind not working from Agent Tunnels view)
  - NX-2837 (Cannot read threshold summary: incompatible operation in Thresholds view on rack)
  - NX-2838 (Agent metrics to read Windows Registry entries)
  - NX-2850 (Tab keyboard button not working in script editor in WEB UI)
  - NX-2853 (Add option to merge separators for agent external table)
  - NX-2856 (Business service availability report crash server process)
  - NX-2859 (Add internal table with notification channels)

I likely fixed it in upcoming 5.2.7, but test on real device is required.

We have published patch release 6 fro NetXMS 5.2 It fixes DB upgrade issue for PostgreSQL as well as few other less important issues.

Full change log since release 5.2.5:
- Added database manager lock to prevent starting background upgrade when it is already running
- Post-upgrade welcome page can be disabled in client preferences
- Fixed bug in handling of duplicate samples sent by agent in cache mode
- Fixed bug in upgrading PostgreSQL database with duplicate data records
- Fixed issues:
  - NX-2832 (Filter not working in Business Service Checks view)
  - NX-2833 (Line chart stacking does not work for data sources with Invert values on)
  - NX-2835 (Add option to disable welcome screen message)

Yes, we work with StarWind since 2017.

Best regards,
Victor

We just published patch release 5 for version 5.2. It wasn't actually planned, but we fix quite a lot of issues while developing version 5.3, so we decided that it is worth backporting some of those fixes and improvements into 5.2 branch.

Full change log since previous patch release:
- Agent configuration file options can be passed from command line
- New AIX agent metrics: System.Memory.Physical.Client, System.Memory.Physical.ClientPerc, System.Memory.Physical.Computational, System.Memory.Physical.ComputationalPerc
- User ACL reports generated in XLSX format
- Improved driver for Ubiquity AirMax devices
- Running configuration logged after log rotation
- Configurable maximum size for cached routing tables
- New metric USB.ConnectedCount in WinNT subagent
- Fixed incorrect object status calculation after restarting server with resolved alarms
- Fixed bug in network map display in dashboard widget
- Fixed bug in calculation of physical CPU usage on AIX
- Fixed missing agent database table file\_integrity
- Fixed process matching by user name in process information metrics in WinNT subagent
- Fixed incorrect detection of Windows Server 2025
- Fixed issues:
   - NX-2794 (Add more info to log message: Potential node x.x.x.x in zone x rejected (IP address is known as cluster resource address))
   - NX-2800 (Agent crash on Windows ARM systems)
   - NX-2819 (DCI data recalculation not working for TimescaleDB)
   - NX-2822 (Alarms not appearing if nxmc has reconnected)
   - NX-2825 (Tables missing PK, this breaks master-master replication (at least in Postgres)) 

We just published patch release 4 for version 5.2. Changes since previous patch release:

- System event is generated when responsible user for object added or removed
- Support for custom OUI database entries
- Fixed bug in loading NXSL "stdlib" script
- Add BOM sequence skip for structured data extractors (including web service extractor)
- Network map fixes: refresh while drag, jumping labels on links and not saved object location after 'align to grid' action
- NXSL function JsonParse sets global variables $jsonErrorMessage, $jsonErrorLine, and $jsonErrorColumn after parsing error
- NXSL function JsonParse accepts optional second argument to control interpretation of integers as floating point numbers
- Configurable data collection scheduling mode that requires connectivity before scheduling DCI for collection
- Fixed issues:
  - NX-2737 (Switch from obsolete SQLite shared cache mode to WAL mode)
  - NX-2783 (Network discovery view shows "Invalid thread access" error on web)
  - NX-2788 (Changing Objects.Nodes.SyncNamesWithDNS requires server restart, while it should not)
  - NX-2790 (Interface Overview view should have node name when interface is under a circuit)
  - NX-2795 (Selection on map is frequently lost)
  - NX-2797 (Hide passwords in server configuration variables view)
  - NX-2802 (Show units for server configuration variables)
  - NX-2806 (Data collection scheduling mode that requires connectivity before scheduling DCI for collection)
  - NX-2807 (EPP export does not escape < and > in XML attributes)

Hi!

Server reads OIDs 1.0.8802.1.1.1.1.2.1.1.1 and 1.0.8802.1.1.1.1.2.1.1.2 to determine PAE state of an interface. Could you please check what is returned there by those switches?
For capability detection server reads OID 1.0.8802.1.1.1.1.1.1.0 and expects value 1. Please check what is returned by Huawei switch for this OID.

Another possible issue could be physical port detection - server only runs PAE status poll on interfaces marked as physical ports. Please check if interfaces are marked correctly as physical ports on Extreme switch.

Best regards,
Victor

Just fixed it, fix ill be in next patch release. Thanks for reporting!

Best regards,
Victor

You can try to set debug level to 7, run instance discovery poll on node, and check server log file for lines with text queryWebService - it may give some hint at what is going on with request.

how do you see this process in "Processes" view for that node?

Are you getting timeout error when you open "routing table" view for Mikrotik device? If so, this likely means bug in a server, as routing tables are cached, and you should get either empty one or last known. Timeout probably indicates deadlock within a server.
On what platform your server is running?

Best regards,
Victor

Hi!

Any chance you can provide me remote SNMP access to some test controller? That would simplify driver development significantly. In any case, could you please point out to relevant MIB files? I will check for differences with what existing driver is using, and decide if it can be adapted or if it should be completely different driver.

Best regards,
Victor

We just published patch release 3 for version 5.2. Changes since previous patch release:

- Default timeout for service checks via netsvc subagent set to 1 second
- Fixed alarm severity text for Grafana API
- Forced plain text web service requests are cached separately
- Fixed issues:
   - NX-2766 (Template macros are not expanded in instance data filed in DCIs Instance discovery)
   - NX-2768 (Changes to VNC properties not logged to audit log)
   - NX-2770 (Object query result view pinning does not work)
   - NX-2777 (No favicon in new Web UI)
   - NX-2781 (Send all parameters of default email notification channel to reporting server)
   - #140 (SQL errors after converting database to TimescaleDB)

We've been participating in GITEX Global in Dubai since 2018, so it's only natural that we want to be part of the first-ever GITEX Europe as well! And so, next week (from 21st to 23rd of May) we will be in Berlin, and if you happen to be there, feel free to visit us! You can find us at Latvia national stand, H4.2-E40.

P.S. Drop us an email or PM me if you'd like to attend and need a complimentary exhibition ticket.

In new REST API (the one loaded as server module) you can do POST on endpoint /v1/objects/{object-id}/set-managed where {object-id} is ID of interface object and content is JSON
{
   "managed":false
}

For old REST API (the one implemented as Java web app) request body will be the same but URL will look like /objects/{object-id}/set-managed

Best regards,
Victor