NetXMS Support Forum

Hi!

I've found a bug in CPU usage calculation code. Fix will be included in 1.2.9 release.

Best regards,
Victor

Скорее всего ето устройство не поддерживает стандартный system MIB. Такая проблема обсуждалась здесь: https://www.netxms.org/forum/general-support/problem-with-the-ethernetbox-from-messpc/. Чтобы сервер мог обнаружить SNMP, добавьте custom attribute snmp.testoid со значением, выставленным в доступный OID.

А core файл создался? Если да, то пришлите пожалуйста stack trace:

# gdb netxmsd core
gdb> bt

Если core нет, то запустите под gdb:

# gdb netxmsd
gdb> run -D3

и после sigfault опять таки

gdb> bt

Quote from: tivanni on September 19, 2013, 05:32:17 PM
An info: how many packets of the 3 packets netxms have to lost for consider ICMP poll failed?

All of them. Logic here is that if any response is received, node is up. If you need to test quality of the link (for example, packet loss percentage), you can use PING sub-agent which can provide you such data.

Best regards,
Victor

Hi!

You can put threshold on DCIs you have created for interfaces, and process generated event to produce notification you need. You can use this tutorial as an example: http://wiki.netxms.org/wiki/How_to_send_e-mail_notification_when_disk_space_is_low.

Best regards,
Victor

Hi!

It should work that way (with semicolon). What is in server log when you are sending SMS to multiple numbers?

Best regards,
Victor

А какое значение isSNMP в разделе Capabilities на закладке Overview для этого узла?

Quote from: Nikk on September 19, 2013, 03:07:53 PM
So, if i'm right, then with appropriate knownledge, I basically can get any information I want?

Yes, exactly.

Best regards,
Victor

Hi!

We are working on sub-agent called DBQUERY which will eventually replace ODBCQUERY sub-agent. It will allow you to do table query and retrieve it as table DCI in NetXMS. Some limited version will be included in 1.2.9.

Best regards,
Victor

Hi!

Transformation scripts are very flexible - main limitation is what data you can access. First of all, all scripts run in strict sandbox environment, and cannot access any data outside of NetXMS server process. It is also possible to restrict script's access to data of nodes other then current (see http://wiki.netxms.org/wiki/SG:Security_Issues for explanation).

You can collect any data with NetXMS, it just matter of efforts you'll need to get them. There are few possible ways to get data:

1. SNMP - it is here out of the box, you can collect anything provided by SNMP. You can freely add new MIBs, but this is not necessary for data collection - only for MIB tree visualization for seemlier configuration.

2. agent extensions - you can develop your own sub-agents to provide application or device specific data (currently you have to do this in C, but later this year we will introduce support for sub-agents written in Java); if you can get your data with command line tools, you can integrate them into agent as external parameters or external parameter providers.

3. Application agents - we have small and simple "application agent" library which can be embedded into your application and provide internal data to NetXMS agenet.

4. SQL database - with ODBCQUERY sub-agent you can read data directly from SQL database.


We are trying to keep NetXMS as secure as possible. You have very flexible access right control; all access checks done in server, not in GUI nor API, so it's not possible to avoid it by using API directly for example. NetXMS uses own communication protocol, which is encrypted using AES-256 by default. Encryption scheme is pretty much standard - on client connect, server sends it's public RSA key to the client; client generates AES session key and sends it to server encrypted with server's public key. Same protocol used also for server-agent communications.

Best regards,
Victor

Hi!

Currently it's hard-coded value of 3 packets. I can make it configurable.

Best regards,
Victor

Hi!

Please post content of /proc/stat from problematic node.

Best regards,
Victor

Сейчас без создания отдельных зон для каждого такого узла это сделать нельзя. Я подумаю, что можно будет сделать.

Я примерно представляю, в какой ситуации могла возникнуть такая ошибка, хотя повторить мне не удалось. На всякий случай изменил генерацию имен для prepared statements, в 1.2.9 это исправление войдет.

Да, можно. Процедура такая:

1. Остановите NetXMS сервер
2. В базе NetXMS выполните запрос

DELETE FROM users WHERE id=0;

3. Запустите сервер.

После этого можно будет зайти пользователем admin  с паролем netxms.