NetXMS Support Forum

Hi!

You have two options, and both requires script creation. One is to use nxshell script - there is an example of how to disable SNMP polling: http://wiki.netxms.org/wiki/Using_nxshell_to_automate_bulk_operations#Disable_SNMP_polling_for_node. Other option is to use NXSL script on server - you can use node enumeration example: http://wiki.netxms.org/wiki/Script_Example:_Enumerate_All_Nodes and add calls to methods enableAgent, enableConfigurationPolling, enableIcmp, enableSnmp, enableStatusPolling, enableTopologyPolling on node objects as needed. Each of these methods accepts true or false as parameter. For example, to disable agent for node you can call

n->enableAgent(false);

assuming that you have reference to node object in a variable n.

Best regards,
Victor

Исправил эту проблему, изменения войдут в релиз 1.2.15.

Hi!

Currently there are no other way to show dashboard without giving read access to nodes where data is collected, so your approach is absolutely right.

As for the script, it's actually a bug - in test mode $dci variable is not set. I'll fix it in the next release.

Best regards,
Victor

Actually it's an error in documentation. For server commands macros are expanded on server side, and you should use event processing macros (https://www.netxms.org/documentation/adminguide/event-processing.html#macros-for-event-processing).

It could be either condition object ID or name - if number is given server will interpret it as ID otherwise as name. If condition name is a number then you should use object ID

It's a bug that "Trusted nodes" is not shown for condition object. Just fixed it in development branch. As a workaround you can use legacy console to edit condition's trusted nodes list.

Best regards,
Victor

Hi!

It seems that rule named "Show alarm when node is down" has any event as a source (probably SYS_NODE_DOWN that was there originally was accidentally deleted) - that way it generates alarm on every event (including node status change events), thus triggering infinite loop of alarm creation. Add SYS_NODE_DOWN back and everything should be fine.

Best regards,
Victor

Hi!

Just test it, and it works as expected. Can you please show full script?

Best regards,
Victor

Можно, хотя и несколько нетривиальным методом. Команды, которую нужно выполнить, настроить как ExternalParameter. При наступлении нужного события запустить NXSL скрипт, в котором при помощи AgentReadParameter получить результат выполнения команды.

Lines 736 and 763 in file src/server/core/objtools.cpp should be


  DBBind(statment, 1, DB_SQLTYPE_INTEGER, dwToolId);


instead of


  DBBind(statment, bindID++, DB_SQLTYPE_INTEGER, dwToolId);


Best regards,
Victor

Hi!

Yes, instance discovery ignores proxy node settings. I register it as a bug.

Best regards,
Victor

I only see event in exported file, not event processing policy rules. Did you add rules for export?

Best regards,
Victor

Hi!

This bug already fixed in 1.2.15. It's not possible to fix it in 1.2.14 without code change. I can provide a patch for 1.2.14 if you are building from sources.

Best regards,
Victor

Discovery в удаленных зонах есть в планах. В принципе почти все для этого есть - осталось только сам процесс discovery подправить.

Не критично - просто эти удаленные объекты будут висеть в памяти до рестарта сервера.

А можно результат SNMP walk по .1.3.6.1.2.1.2?