NetXMS Support Forum

Hi!

I was under impression that you are using NetXMS agent on server with ppp interfaces. Can you test what interface list returned by SNMP agent using SNMP walk? Like this:

nxsnmpwalk -c community server_address .1.3.6.1.2.1.2.2.1.2

I suppose that it is a bug in Linux snmpd which causes to return all previous instances of ppp interface along with current one.

Another problem is changing interface index. I take a quick look inside ppp module in Linux kernel, and it looks like it assigns new interface index on each ppp connection setup. Because when you collect interface data via SNMP the only option to identify interface is by index, you cannot create simple DCI which will always work, because you'll need to change SNMP OID when interface index changes.

The easiest way to solve this would be to install NetXMS agent on this Linux server - then you'll be able to gather data by interface name. Event if it's some kind of appliance installing NetXMS agent is most likely possible if you can access it's file system and execute commands.

Best regards,
Victor

Hi!

We don't get much patches, so there are no established workflow I suggest that small patches for fixing specific bugs should be submitted via our bugtracker. Regarding large changes please contact me first and we will discuss better way.

We don't have plan to migrate to public repositories. As we are trying to build business around NetXMS (which will allow as to keep working on open source project), we have some closed source extensions and need repository we have full control of.
Our current repository is subversion, but we plan to migrate to git soon. I'll provide you with necessary access as soon as you'll need it.

Best regards,
Victor

Это надо делать на том агенте, который будет подключаться к базе. На скриншоте отсутствует начало секции ORACLE, возможно в этом проблема. Для диагностики можно запустить агента в дебаг режиме (с ключом -D9), тогда будет видно, подключается он к базе или нет (только перед этим обязательно поставить лог в файл).

Hi!

nxaction is a tool intended for connecting to agent and executing action on agent (defined in nxagentd.conf using Action parameter). For executing server action simpliest way is to use nxevent for sending custom event and executing action you need in event processing policy as usual.

Best regards,
Victor

Hi!

It should be like this:


WMI.Query(root\CIMV2\, Select * from Win32_NTLogEvent Where Logfile = 'System' and EventType = '3' and EventCode = '1020',%%count%%)


Best regards,
Victor

Hi!

Just got to issue 303. It turns out that limit for events and alarms messages was already raised to 2000. I do tests with similar security log events. Text is there, but looks like list control has some limitations as well - text is truncated when you look at events or alarms in tabular form. However, if you open alarm detail or copy appropriate record to clipboard and paste into text editor you'll see full message text.

Best regards,
Victor

Hi!

Strange that old interfaces remains in interface list. Normally it should be deleted and new interface object with same name but different index created. Can you post output of

nxget -l <server_ip> Net.InterfaceList

before and after PPP session comes down and up again. It should not affect data collection because you can use interface names instead of indexes in all Net.Interface parameters.

Best regards,
Victor

%n - это имя обьекта в NetXMS. Вы уверены что оно совпадает с DNS именем и что с разрешением DNS имен на сервере все в порядке? Для надежности можно использовать %a - IP адрес ноды.

Скорее может чем нет Для NetXMS есть субагент для мониторинга Oracle (http://wiki.netxms.org/wiki/Subagent:Oracle), он дает кое-какую информацию. Мы его можем легко расширить, проблема скорее в том что мы не знаем, что именно надо мониторить в Oracle в реальной жизни. Так что пробуйте, по отзывам будем улучшать агента

Hi!

You have two options here. If you want to ignore all events for particular node(s), the easiest way is to add rule at the top of the policy with these nodes in filter and action set to "stop event processing" (see screenshot "rule_1"). To ignore node(s) in just one rule, you can add it to the filter and check "Inverted rule" on "Source" page in rule properties (screenshot "rule_2"). This way filter will match events from all nodes except those specified.

Best regards,
Victor

Hi!

Currently the only way is to turn on debug on server (level 5 or higher) and check server's log. We have a feature request for that (https://www.radensolutions.com/chiliproject/issues/97) and eventually it will be implemented.

Best regards,
Victor

Hi!

In fact, instance discovery using tables is not working yet. You can use agent list as discovery method and use list FileSystem.MountPoints.

Best regards,
Victor

Hi!

You can set "required poll count" for node interface(s) to desired number of polls. NetXMS will consider interface down only after given number of unsuccessful polls, and so entire node will be considered down only then.

Best regards,
Victor

Strange. Should work. Try to do

make clean
make
make install

just in case and try again. After that, try to start server with debug level 6 and check log for possible errors related to encryption.

Best regards,
Victor

Seems to be an instance discovery issue. I'll try to reproduce it on my test system.

Best regards,
Victor