NetXMS Support Forum

Hi,

you can use NXSL script and function AgentReadList like this:


AgentReadList($node, "SSH.Command(mikrotik_ip,login,password,/system package update install)");


where $node is a node with SSH subagent installed. Of course you can use string concatenation to insert proper credentials and device IP address. For example, if variable $node points to NetXMS server node and variable mt to Mikrotik device, you can do


AgentReadList($node, "SSH.Command(" . mt->ipAddr . ",login,password,/system package update install)");


Best regards,
Victor

Quote from: Egert143 on May 27, 2019, 10:11:15 AM
I would like to ask one additional question, but not sure if i should start new thread for that. Is there possibility to have a "safe / know" list of devices per "site / subnet" for example known good mac + ip combo and when new device appears with network discovery it will get flagged?

Yes, it's possible with some scripting. Rough idea is to create mapping table with good MACs, setup script action to SYS_NODE_ADDED event, and in that script check if one of node's MAC addresses is in that table.

Best regards,
Victor

Hi,

predefined graphs and dashboards configurations are completely independent now. Adding predefined graph to dashboard seems like a good idea, I'll add it as feature request.

Best regards,
Victor

This was the bug in 2.2.14. Actually nxdbmgr check from 2.2.15.2 should have fixed that issue, you hadn't to reinstall everything.

Best regards,
Victor

Update: in release 2.2.15 size limit for instance field increased to 1023 characters.

Best regards,
Victor

Hi all!

NetXMS version 2.2.15 is available. It is mostly hotfix release for issues found in version 2.2.14. Full change log is following:

- Improved driver for Cisco Small Business switches
- Fixed issues:
        NX-1609 (Duplicate server configuration parameters)
        NX-1627 (Proxy node should not ping through itself)
        NX-1637 (Unrecognized Cisco SG Switch)
        NX-1639 (NXSL persistent storage entries not saved in database)
        NX-1644 (DB gets corrupted when adding container/nodes)
        NX-1645 (Broken UI element layout in table DCI "General" property page)

It is highly recommended to run nxdbmgr check after upgrading binaries and before first server start.

Best regards,
Victor

We just discovered critical bug in version 2.2.14 that can cause container relation information not to be saved in database (registered as https://track.radensolutions.com/issue/NX-1644). It is recommended not to upgrade to 2.2.14 or do not do any changes to container structure until upgrade to 2.2.15. We are working on release 2.2.15 with a fix, it should be available soon.

Best regards,
Victor

Добрый день!

Если на ноде не осталось данных по "peer interface" (это видно в информации по интерфейсу), то не получится. Информация об истории подключений не сохраняется. У нас есть запрос на такой функционал, в ближайших релизах скорее всего появится.

wget http://packages.netxms.org/netxms-release_1.2_all.deb
dpkg -i netxms-release_1.2_all.deb
apt update
apt install netxms-server netxms-dbdrv-pgsql

(замените netxms-dbdrv-pgsql на другой драйвер если используется не PostgreSQL)

потом надо создать /etc/netxmsd.conf и /etc/nxagentd.conf, проинициализировать базу командой

nxdbmgr init

и можно стартовать сервер.

Здесь понадобится два скрипта. Делается примерно так:

1. Делаем список HTTPS.Certificates

ExternalList = HTTPS.Certificates:script1.cmd


Скрипт должен выдавать список сертификатов, каждый в своей строчке.

2. Изменяем параметр HTTPS.CertificateExpireDate чтобы он принимал имя сертификата:

ExternalParameter = HTTPS.CertificateExpireDate(*): powershell.exe -file "\\server\Scripts\test.ps1" -certname $1


Насколько я знаю power shell можно будет в скрипте получить имя сертификата в переменной certname. Дальше в скрипте все как сейчас.

3. Делаем DCI на сервере:
Параметр

HTTPS.CertificateExpireDate({instance})


Instance discovery method: agent list
Имя списка: HTTPS.Certificates

Try to start agent with debug level 6 or higher, and check what is going on with incoming connections. Make sure that agent logging is set to file.

Best regards,
Victor

Похоже это следствие ошибки в ReadPersistentStorage - если записи с заданным ключом нет, то возвращается пустая строка, а не NULL. Правильным вариантом кода будет

ReadPersistentStorage($event->dciId . "_0x" . right(d2x($node->id), 8, "0"));

Предварительно можно еще проверить что $event->dciId не NULL.

Parameter itself is ok, there is 1023 character limit. Problem is with instance value where actual OID will not fit. It is fairly easy to extend instance length limit to 1023 characters as well, I can do that for next release.

Best regards,
Victor

А точно работает? ReadPersistentStorage на делает подстановку, т.е. будет искаться ключ %<dciId>_%i, а не скажем 127_313.

Можно помимо генерации аларма делать запись с таким-же ключом в persistent storage, и в правиле отсылки почты по SYS_THRESHOLD_REARMED проверять наличие такого ключа, и только если он есть выполнять отсылку.