NetXMS Support Forum

Hi!

You can install test agent on same machine, using the following procedure:

1. Stop netxms agent service
2. Remove agent service by running nxagentd.exe -R
3. Install test agent into separate directory. This is most important! You will have to copy nxagentd.conf.

To recover after test agent uninstall, do the following:

1. Ensure that uninstall has removed agent service
2. Install agent service by running nxagentd -I -c <full_path_to_nxagentd.conf>

Best regards,
Victor

A kakaja imenno versija Windows?

It general, it is not possible to distinguish from outside if node was shut down or cable was disconnected. The only situation where it can be distinguished that I can think of is following:

Assumptions:

1. If node is off (shutdown), it still have power and link on NIC is up
2. Node is connected to managed switch and port is known

Then you can check port state on switch - if it is down, then most likely cable is disconnected (or node was completely powered off). If it is up, but node is not responding, then node was shutdown.

Also, if node in question is a server with management board (like HP iLO), you can double check server state via management board.

Best regards,
Victor

Hi!

You mean NetXMS server itself or monitored node? For monitored node, server will generate SYS_NODE_DOWN event if node is unreachable. You can create a rule in event processing policy to send email, generate alarm, etc. as a reaction to this event.

Best regards,
Victor

Hi!

Sorry, I forgot to add it to source distribution. Script is attached. It is very simple perl script which uses SSSU to get information from EVA, so you can use it on any machine with Perl and SSSU. Add the following line to your nxagentd.conf:

ExternalParametersProvider = <path>/hp_eva_collect_data.pl

and restart agent. After configuration poll, you should see additional parameters for each controller, disk shelf, and disk in EVA.

Best regards,
Victor

No, I do it today. Version 1.1.7 will include this feature. If it is needed right now, I van make cudtom build of agent.

Best regards,
Victor

Hi!

I have added special option to get number of rows in WMI query result set - if you give %%count%% as property name, agent will return number of rows.

Best regards,
Victor

Hello!

Sorry for delay with that. I do some additional testing, but was unable to reproduce this issue, even by installing event sources without english resources. Agent either give different error or renders text as it should. I have made small changes in log watch subagent, and add logging of message file name after EvtFormatMessage failure, so at least it will be possible to see what application causing this issue. Patched agent version can be downloaded here: https://www.netxms.org/download/dev/nxagent-1.1.6.2-x64.exe.

Best regards,
Victor

Hi!

Done in 1.1.6 client. In the next release, I also plan to add possibility to configure different sounds for different alarm priorities.

Btw, how are overall impression with Android client? Is it useful, what functions are most desired in mobile client?

Best regards,
Victor

Zdes' glavnaja problema v tom, chto ne suschestvuet universal'nogo algoritma, po kotoromu mozno opredelit', proizoshel li sbros schetchika, ili eto normal'noe izmenenie znachenija. Edinstvennij poka variant, kotorij ja vizu - eto dat' vozmoznost' dlja kazdogo DCI zadavat' maksimal'noe teoreticheski vozmoznoe znachenie del'ti - togda server budet ignorirovat' takie piki.

Hi!

Yes, it's ok. That way it's far less chance that problem will be forgotten

Best regards,
Victor

Hi!

Java console is only compatible with server version 1.1.x. I recommend you to upgrade to server version 1.1.6.

Best regards,
Victor

Da, dejstvitel'no. Ja vecherom vilozu fail.

Hi!

Currently there are no way to detect alarm ack/termination by operator and do some actions. It's a good idea to add a hook for that. I'll think how it can be implemented.

Best regards,
Victor

Hi!

Acknowledged alarm will remain in list of active alarms, but in "acknowledged" state. Terminated alarm will be removed from active alarm list. Main idea behind this is to provide mechanism for operators to inform each other that problem do exist, but somebody already aware of it and working on it.

Best regards,
Victor