NetXMS Support Forum

Hello,

Would it be possible to extend/add System.CPU.Count() metric provided by subagents/linux with parametrized one that would enable count of online/offline CPUs in the system.

Example system with 24 cores with 2 cores online:
...:# lscpu --online --extended
CPU NODE SOCKET CORE L1d:L1i:L2:L3 ONLINE    MAXMHZ    MINMHZ       MHZ
 12    0      0    0 0:0:0:0          yes 3200.0000 1200.0000 1197.3630
 13    0      0    1 1:1:1:0          yes 3200.0000 1200.0000 1197.3240

...:# lscpu --offline --extended
CPU NODE SOCKET CORE L1d:L1i:L2:L3 ONLINE    MAXMHZ    MINMHZ       MHZ
  0    0      0    0 0:0:0:0           no 3200.0000 1200.0000 1197.3800
  1    0      0    1 1:1:1:0           no 3200.0000 1200.0000 1197.4100
  2    0      0    2 2:2:2:0           no 3200.0000 1200.0000 1197.3630
  3    0      0    3 3:3:3:0           no 3200.0000 1200.0000 1200.0000
  4    0      0    4 4:4:4:0           no 3200.0000 1200.0000 1197.8280
  5    0      0    5 5:5:5:0           no 3200.0000 1200.0000 1197.4139
  6    1      1    6 8:8:8:1           no 3200.0000 1200.0000 1197.3640
  7    1      1    7 9:9:9:1           no 3200.0000 1200.0000 1197.5330
  8    1      1    8 10:10:10:1        no 3200.0000 1200.0000 1197.4250
  9    1      1    9 11:11:11:1        no 3200.0000 1200.0000 1198.4170
 10    1      1   10 12:12:12:1        no 3200.0000 1200.0000 1197.3580
 11    1      1   11 13:13:13:1        no 3200.0000 1200.0000 1197.4180
 14    0      0    2 2:2:2:0           no 3200.0000 1200.0000 1197.5649
 15    0      0    3 3:3:3:0           no 3200.0000 1200.0000 1197.3621
 16    0      0    4 4:4:4:0           no 3200.0000 1200.0000 1200.0000
 17    0      0    5 5:5:5:0           no 3200.0000 1200.0000 1197.4301
 18    1      1    6 8:8:8:1           no 3200.0000 1200.0000 1200.0000
 19    1      1    7 9:9:9:1           no 3200.0000 1200.0000 1197.5850
 20    1      1    8 10:10:10:1        no 3200.0000 1200.0000 1197.5100
 21    1      1    9 11:11:11:1        no 3200.0000 1200.0000 1197.5310
 22    1      1   10 12:12:12:1        no 3200.0000 1200.0000 1197.5060
 23    1      1   11 13:13:13:1        no 3200.0000 1200.0000 1197.3719
Metric is provided via sysfs:

# cat /sys/devices/system/cpu/present
0-23

# cat /sys/devices/system/cpu/online
12-13

Etc:
- on this exact system System.CPU.Count will return: 14 for number of CPU's.

thank you,

Br, Ales

Hi,

I would expect that these devices support SNMP. If this is so, you need to enable snmp on the device and run configuration poll in NetXMS to check if SNMP capablitiy is detected (it's visible on node's Overview page). When this is done, you can use MIB explorer to walk available OIDs and add data collection items.

The other possible approach is to use ssh - in node properties you can specify ssh password or key. Or, if you have many devices, credentials can be specified in Configuration -> Network credentials - these will be auto-probed. If ssh credentials are ok, on configuration poll SSH capability should become "YES".

And, finally, if your device runs under linux, you can try generic NetXMS agents - but this depends on platform/cpu and not necessarily would work

Hello,
I'm new to NetXMS.
Someone can tell me if there are templates / method to monitor Storage SAN?

Any suggestion will be appreciated

Hello,

Q: Was it working on older versions or this is new setup?
A: Unknown. This is a new/clean setup. Specific is tunneled agent setup. And agent uptime monitoring (which pointed out issue).

Q: What exact version of openssl package you are using?
A: Package: libssl3, Version: 3.0.11-1~deb12u2, APT-Sources: http://deb.debian.org/debian bookworm/main amd64 Packages

Br, Ales

Tried clean test server install with just the node that do not work and it works on test server.
Tried renaming and then recreate the node that does not work and now snmp v3 works on the new node on production server.
Odd issue that only snmp v3 is not working on some old nodes.

I will recreate all nodes to be safe.

Agent crashes within OpenSSL library (specifically within EVP_PKEY_CTX_new) and I do not see any obvious problems with calling code. This part was not changed for quite some time. Was it working on older versions or this is new setup? What exact version of openssl package you are using?

Best regards,
Victor

nxshell is integration tool, that allows python code to interact with netxms server. For Windows GUI client you need to install this: https://netxms.com/download/releases/4.5/netxms-client-4.5.6-x64.exe

Current Windows WEB UI installer includes web api as well. It can be reached at http://127.0.0.1:4788/api/
War file itself is located in C:\NetXMS-WebUI\jetty-base\webapps\api.war

Yes, debug log on level 6 would be beneficial. Couple of options
 - use some netxms agent in proxy mode and take the log of it
- use some test installation of netxms server that has just a couple of SNMP nodes (it's also worth to unmanage the node that represents server itself to minimize logged stuff)

Hi,

Any known issues with connecting to snmp v3 agents?
Have odd issue where some nodes work and some don't with the exact same auth configuration.

The node that netxms can't connect to agent with v3 is verified that the agent is running and v3 auth information works in other snmp explorer software.

The node can connect with v2c if enabled, just v3 having issues.

Tried debug logs but since it is the server very hard with spammed debug log, set snmp.discovery to level 6 and it shows SnmpCheckV3CommSettings trying then fail.

When i do full config poll on a node that V3 works on there is no SnmpCheckV3CommSettings in logs it just connects and polls.

The nodes that V3 work on and the nodes that does not have exact same config no question since the snmp is set with nxshell script on all nodes in the container.