NetXMS Support Forum

Hello!

No, NetXMS can work only with it's own web server. Anyway, web interface is for read-only access and is not required. You can use only management console and do not install web interface.

Best regards,
Victor

В текущей версии фильтровать по severity в syslog'e нельзя. Добавлю в следующей версии (добавил в bugtracker под номером 263).

You can configure these tools via Control Panel -> Object Tools.
For example, to configure telnet tool under Tools->Connect, add new tool of type Command with name &Connect->&Telnet. In the "Command" field, enter telnet %OBJECT_IP_ADDR%, and add Everyone (or specific users/groups) to access list.

Best regards,
Victor

Hello!

Attached is a simple program that will print size of address space available to applications in megabytes. It will show 2047 for 2GB mode, and something near 3072 for 3GB mode. I can add this as agent's parameter in next release.

Best regards,
Victor

А пользователь тот-же при запуске скрипта руками, под которым агент работает? Единственно что приходит в голову - разница в environment и/или другой пользователь при запуске из-под агента и руками. Кстати, а если скрипт этот скажем через cron запустить, он работает?

Hello!

You cannot change when SYS_NODE_DOWN sent. However, you can process it depending on other node's status using situations. For example, you have to nodes: node_A and node_B, and wish to track their status, you can do the following:

1. In console, go to View -> Situations;
2. Create new situation called NodeDown (you can choose any name of course);
3. Open event processing policy
4. Create new or change default rules for SYS_NODE_DOWN and SYS_NODE_UP events: in situation column, select Edit, than check "Update situation if current rule match"; select situation NodeDown; as situation instance, enter %n (this will create separate instance for each node, each instance identified by node name); in the attribute list, add attribute called "status" (again, you van choose any name here), and as value enter 1 for SYS_NODE_DOWN event, and 0 for SYS_NODE_UP.

Now we have defined situation, which will have separate instance for each node, and for each instance wil will have attribute "status" set to 1 if appropriate node is down, and to 0 if it is not.

Now you can use this information in filtering scripts by accessing the situation:

s = FindSituation("NodeDown", "node_name");

and looking at situation's attribute "status":

s->status

For example, if I wish that certain event will be processed only if both node_A and node_B are down, I can use the following filtering script:


sub main()
{
   s1 = FindSituation("NodeDown", "node_A");
   isDownA = (s1 != null) ? s1->status : 0;
   s2 = FindSituation("NodeDown", "node_B");
   isDownB = (s2 != null) ? s2->status : 0;
   return s1 && s2;
}


Construction (s1 != null) ? s1->status : 0 is needed because FindSituation will return null if given instance does not exist. In our example, if situation's instance for node does not exist, we assume that this node is up (status attribute equals 0).

Hope this helps!

Best regards,
Victor

Похоже дальше надо вставлять отладочную печать в скрипт. Можно добавить в скрипт, чтобы он выводил результат не только на stdout, но и в какой-нибудь файл? Тогда сразу будет понятно, в какую сторону дальше копоть.

When you start agent as standalone program, not as system service, you should not use -d option. So, to debug, you should use command

"C:\Program Files (x86)\NetXMS\bin\nxagentd.exe" -D -c "C:\Program Files (x86)\NetXMS\etc\nxagentd.conf" -M "192.168.0.1"

Best regards,
Victor

Hi!

Yes, it's the same bug. I have added it to bug tracker (number 262) - it was not registered before.
The only way to add columns manually is to modify database (you can use existing table tools as an example) and then restart server.

Best regards,
Victor

Quote from: oleg on August 10, 2009, 07:09:55 PM
Добрый день!
Такой вопрос.
А как можно парсить логи имя которых не постоянно, а типа log090810.log? На следующий день соответственно log090811.log.

В имени файла можно использовать макросы:

%a   Abbreviated weekday name
%A   Full weekday name
%b   Abbreviated month name
%B   Full month name
%c   Date and time representation
%d   Day of the month (01-31)
%H   Hour in 24h format (00-23)
%I   Hour in 12h format (01-12)
%j   Day of the year (001-366)
%m   Month as a decimal number (01-12)
%M   Minute (00-59)
%p   AM or PM designation
%S   Second (00-61)
%U   Week number with the first Sunday as the first day of week one (00-53)
%w   Weekday as a decimal number with Sunday as 0 (0-6)
%W   Week number with the first Monday as the first day of week one (00-53)
%x   Date representation
%X   Time representation
%y   Year, last two digits (00-99)
%Y   Year
%Z   Timezone name or abbreviation
%%   A % sign

В нашем случае это будет log%y%m%d.log

Pohoze na bug v servere. Esli vozmozno, v moment zavisanija prodelat' sledujuschee:

1. na servere zapustit' komandu nxadm -i
2. V pojavivshejsja komandnoj stroke vidat' sledujuschie komandi:

show pollers
show queues
show mutex
show sessions
show watchdog
show flags

3. Prislat' vivod etih komand na [email protected]

Budem razbiratsja.

Hello!

Mostly it's a traces of planned but not implemented functionality. Event menu does not working at all, as well as export menu. Print will work only in graphs.
We are actively working on completely new management console, which should become usable in a few month, so I don't want to spent time on improving existing console.

Best regards,
Victor

You can take a look at skeleton subagent - src/agent/subagents/skeleton. It can be used as a template for creating your own subagent.
For restarting specific services - you do not need any development for this, it can be done via configuration. You can define actions in agent and call them either from event processing policy or via object tools.
For storing additional information on server, you can either create your own additional table in database, or use object's custom attributes. Anyway, you will need to write small server module to handle this.

Best regards,
Victor

Hello!

Unfortunately, development documentation almost absent. There are document which describes communication protocol, but it covers only message format and small subset of commands. Also, there are some notes in doc/internal directory in source tree. And there are some comments in source code.

Of course, you can ask questions here, and I'll do my best to answer them. Maybe it will become a foundation for Development FAQ You can also consider ordering custom development from Raden Solutions (http://www.radensolutions.com).

Best regards,
Victor

Mi delaem sejchas polnostju novuju konsol', tam budet normal'nij log viewer s fil'trami. Ja skoro vilozu alpha-versiju. Sdelat' eto v staroj konsoli zajmet kuchu vremeni, i uze ne hochetsja ego tratit' na ustarevshij komponent.