NetXMS Support Forum

Полный функционал сейчас только в Java API. Старый C API (libnxcl) из 2.0 убрали, сейчас есть новая C++ библиотека (libnxclient), но набор функций там пока минимальный. Можно ее по необходимости расширять.

Добавление данных прямо в базу работать не будет, сервер кэширует данные и не ожидает их изменений в базе.

transformation script не подходит, он не вызовется если входных данных нет. Для агентских DCI подход тот-же, только заменить SNMPGet на AgentReadParameter. В будущем можно пожалуй добавить опцию в DCI для вызова transformation script в случае ошибки с $1 выставленным в null скажем - это упростит настройку множества таких параметров.

Ставить знак вопроса. Первый параметр подставится в первый знак вопроса, второй во второй, и т.д. Например:

SELECT STUFF ( (SELECT top 5 ';' + objid + ',' + action + ',' + CONVERT(varchar(10),DATEDIFF(s,'1970-01-01', DATEADD(second, DATEDIFF(second, GETDATE(), GETUTCDATE()), date))) FROM .... WHERE ... condition_column=?

Нет, они никак не связаны. Можно добавить как feature request, технически сделать не сложно.

Можно сделать script DCI, который будет делать сбор данных и подставлять константу в случае ошибки. Для SNMP может выглядеть например так:


snmp = CreateSNMPTransport($node);
if (snmp == null)
   return -1; // error indicator
v = SNMPGet(snmp, ".1.3.6.1.2.1.1.1.0");  // actual OID can be read from script parameter
if (v == null)
   return -1; // error indicator
return v->value;


Hi,

I suppose you mean agent? Then you can get one of older versions from archive: https://www.netxms.org/download/archive/agent_packages/. You can try 1.1.10, if it will not work, then 1.0.13.

Best regards,
Victor

Hi,

we just updated Windows installers and source package with new build that contain a fix. File names are the same, but after install you should see build number 8036 for server binaries. Debian repository will be updated in a few hours.

Best regards,
Victor

Hi,

assigning cluster resources with a script will definitely allow more possibilities for cluster monitoring. I'll try to add it in one of next versions.

Best regards,
Victor

Hi,

it seems that I just have to change base OID for SNMP walk. I've added issue to pur bug tracker: https://dev.raden.solutions/issues/878.

Best regards,
Victor

Hi,

I was able to reproduce this problem on my test system. I'll provide fix soon.

Best regards,
Victor

Quote from: brettmilfos on August 11, 2015, 12:53:50 AM
I have encountered one issue. While upgrading from 2.0M5 all the mibs from var/mibs dis-appeared. This is on win2k8r2.

Restored mibs from a backup and re-built. But now when I run the 32bit console on win7 I get the following error after logging in:
"Cannot load MIB file from server: File I/O operation failed"

MIB locations was changed in 2.0-RC1. Source MIBs now are in <install_path>\share\mibs and compiled MIB file in <install_path>\var. Installer was supposed to move any non-standard MIBs from var\mibs to share\mibs. If you compile MIBs manually, you should set output file to be <install_path>\var\netxms.mib.

Quote from: brettmilfos on August 11, 2015, 12:53:50 AM
Also, do not know if this is related to the upgrade, before upgrading I stopped the Core, Agent & WebGUI services, but there was still a lock on the MySQL DB which prevented DB upgrade.

That means that core service didn't stop normally and was forcibly terminated by service manager. It's not related to upgrade, but rather indicates some problem with previous version.

Best regards,
Victor

Hi,

instead of processing SYS_AGENT_UNREACHABLE event you can collect internal parameter AgentStatus. It returns 0 if agent is connected and 1 if unreachable. Then you can use thresholds as usual to generate alarm only if agent is unreachable for more then specific time period.

Best regards,
Victor

Can you provide more detailed information how exactly false positive looks like (service name, etc.). I tested it on my machine and everything works as expected.

Best regards,
Victor

Hi all!

NetXMS 2.0-RC1 is out! Changes since previous release:

- Fixed LDAP authentication issues with Active Directory
- Fixed character encoding issues in LDAP user data
- Java subagent improved
- Fixed multiple issues with agent side data caching
- Hash maps implemented in NXSL
- Improved array implementation in NXSL
- New NXSL function ArrayToString
- New OS parameters and tables: System.HandleCount, System.OpenFiles, Process.Handles
- Unified macros in all types of object tools
- Server can be switched to case-insensitive login names mode
- Implemented support of "command generates output" option for server commands
- Network device driver for TelcoBridges gateways
- Fixed memory leak in Qtech OLT driver
- Java subagent plugin for Ubiquity/LigoWave device monitoring
- Management console:
   - Default object display mode can be set in map properties
   - Map drill-down can be enabled in dashboard map elements
   - Input fields in object tools
   - Improved script editor for script library
   - Interface speed displayed in interface object details and in "Interfaces" tab
   - Symbolic name of interface type displayed when known
- Fixed issues: #38, #55, #741, #818, #829, #830, #836, #838, #840, #843, #844, #845, #846, #848, #850, #853, #855, #856, #857, #862, #866

Best regards,
Victor

It is not possible now, but I think it could be quite useful to have template graphs (and possibly dashboards) preconfigured and accessible for node in a way similar to object tools. So far I see easiest way to implement it is to add option "save as template" for graph - it will save all settings except actual node name, and when called for node will show same parameters for selected node. It should not be too complicated to implement. For those who wants to track progress: https://dev.raden.solutions/issues/876.

Best regards,
Victor