NetXMS Support Forum

I have found the problem with classpath setting in JMX plugin. Try to replace jmx.jar on your system with this version: https://cloud.radensolutions.com/s/WDCCtmPKHXGZwyo.
If you installed agent on Linux from packages it will be inside /usr/lib/netxms/java.

Best regards,
Victor

Wildcard import implemented in current development branch and will be included in release 5.0 (and 4.5 if it will be publicly released).

Best regards,
Victor

NetXMS version 4.4.5 is just released. It fixes two critical issues: server crash on polling TP-Link devices (bug that was introduced in 4.4.4), and severe performance degradation when monitoring multiple (thousands) SNMP devices via single proxy.

Full change log since previous patch release:

-  Improved SNMP proxy performance under heavy load
-  Added limit on number of nested NXSL VMs (to prevent accidential infinite loop of script execution)
-  Fixed server crash on polling TP-Link switches
-  Fixed bug in dashboard element "status indicator"
-  Fixed bug in status map view
-  Fixed bug in database manager check function
-  Fixed "Failed to register resource" error in web UI
-  Fixed database import/migration to TimescaleDB
-  Java components switched to logback 1.3.13 (fixes CVE-2023-6378)
Fixed issues:
-  NX-2465 (List of saved queries in Tools->Find Object is not updated when query list is altered in Configuration)
-  NX-2479 (Misleading error messages when loading properties for root objects)

Should be as simple as running
nxdbmgr.exe init
from command line. It may ask additional information and will show result in console.

Best regards,
Victor

Rewriting Grafana plugin is in our pipeline, I hope we can make it in January or February.

Best regards,
Victor

We made a fix recently (in 4.4.x) for stalled client sessions from web API - that may fix this issue too.

Best regards,
Victor

Hi,

this (and other) metrics requires JMX server name as an argument. You define JMX servers in agent configuration file (see https://www.netxms.org/documentation/adminguide/jvm-monitoring.html#configuration) Unfortunately, there is no agent list to return all configured JMX servers on that particular agent (we can add one in next release). So for now you cannot use instance discovery if you have multiple JMX servers monitored by one agent.

Best regards,
Victor

Hi!

Try to use server_name\instance_name in DBServer. Also you can try IP address instead of DNS name.
Check that TCP/IP and named pipes enabled for Native Client.

Another alternative is to create ODBC DSN and switch to odbc driver (use DBDriver=odbc instead of DBDriver=mssql). In that scenario you should use ODBC DSN as value for DBServer.

Best regards,
Victor

NetXMS version 4.4.4 is just published. Changes since version 4.4.3:

-  New methods in NXSL class "InetAddress": contains, equals, inRange, sameSubnet
-  Constructor for NXSL class "InetAddress" accepts mask length as second argument
-  Fixed incorrect ICMP polling if ICMP proxy set on node level
-  Improved topology discovery on TP-Link devices
-  Improved driver for DLink devices
-  Added driver for TP-Link devices
-  Added driver for Eltex devices
-  Added driver for Q-tech devices
-  nxencpasswd can read password from terminal
-  GUI clients built with patched version of simple-xml (fixes CVE-2017-1000190)
-  Fixed deadlock after login in legacy web UI

Fixed issues:
-  NX-2431 (Implement agent list Net.IP.RoutingTable for AIX)
-  NX-2478 (Named function parameters does not work for entry points)

NetXMS version 4.4.3 is just published. Changes since previous patch release:

-  Package deployment can be scheduled
-  Server-side macro expansion in package deployment command
-  Use compact JSON format when saving events to database
-  Improved event processing performance
-  Improved NXSL function "random"
-  New event processing macros %d (DCI description), %D (DCI comments), %L (object alias), and %C (object comments)
-  Added driver for FortiGate devices
-  Fixed server crash during execution of delayed EPP action
-  Fixed server crash when processing interfaces with 8 byte MAC address
-  Fixed session disconnect handling in new management client application
-  Fixed bug in physical disk information reading on Windows
-  Fixed bug in SSH key store
-  Improved debug logging
-  Minor fixes and improvements in new management client application
-  Fixed issues:
   -  NX-1063 (Interface icon is incorrect)
   -  NX-2224 (Command history in nxadm)
   -  NX-2446 (Increase timeout for agent tunnel binding)
   -  NX-2463 (Add metric to measure execution time of background queries in dbquery subagent)
   -  NX-2467 (Allow to execute same action multiple times in one EPP rule)
   -  NX-2468 (NetworkService.Status SMTP call to curl_easy_perform failed (56: Command failed: 502))
   -  NX-2469 (Empty "Parameters" line should be interpreted as no arguments in Execute Script)
   -  NX-2471 (Add agent list and table to list physical disks)
   -  NX-2475 (netsvc: ServiceCheck.SMTP() uses VRFY command, which is disabled on most servers)

Hi!

You should be able to use all the % macros in object tools (same macros as in event processing), so you can use custom attribute to define TCP port for example.

Currently there is no control on local port number - system allocates random port, typically from an ephemeral port range. It should not be a problem for desktop client, as this port is used only on loopback interface. FOr web client it could be an issue though, if you use web connect object tool from web client.

Best regards,
Victor

This looks like a bug in the server. Please try upgrading to 4.4.2, if this will not help, please try to run netxmsd under debugger and provide stack trace after crash.

Best regards,
Victor

There is no such option in the driver. However, it is possible to use configuration poll hook to test if node name is "name" and driver is "TELTONIKA", and rename it as desired. Sample script that does this and sets node name to product name:

if ($node->name == 'name' and $node->driver == 'TELTONIKA')
{
   $node->rename($node->productName);
}

Best regards,
Victor


New patch release 4.4.2 is just published. Changes since previous release:

-  Server checks interface speed during status poll and generates event if it changes
-  Improved Cambium device driver
-  Added driver for Hirschmann switches
-  Implemented implicit import for constants in NXSL
-  NXSL implicit import does not add non-referenced functions and constants from imported module
-  Context action "Change expected interface state" implemented in new GUI client
-  Context action "Clone network map" implemented in new GUI client
-  Masked credentials in "Network Credentials" view
-  Fixed bugs in TCP proxy session closure handling (server and agent side)
-  Fixed bug in parsing XML content returned by web service
-  Fixed template apply/remove in new GUI client
-  Fixed server crash when network map uses physical link with non existing rack
-  Fixed audit log writing on object move

Fixed issues:
-  NX-2410 (Notification driver is locked during retry waiting period)
-  NX-2432 (Query interface speed when status poll detects that interface goes up)
-  NX-2441 (Auto-focus on Two-Factor input on WebUI)
-  NX-2442 (Maintenance predefined time)
-  NX-2449 (Unexpected SYS_DUPLICATE_IP_ADDRESS generation)
-  NX-2450 (microhttpd presence is not detected correctly)
-  NX-2451 (GetDCIValue() should return same data type as set in DCI properties)
-  NX-2452 (Agent on Windows returns only one software inventory record when multiple versions of same software are installed)

Hi,

try just
59 * * * 6-7
for second schedule - it should have same effect.

Best regards,
Victor