NetXMS Support Forum

Hi all!

NetXMS 3.8 patch release 3.8.193 is just published. It is must have update for Timescale DB users as it fixes critical compatibility issue with Timescale DB 2.x.
Full change log:

- Fixed Timescale DB 2.x compatibility issues
- New methods isDirectChild, isDirectParent, isChild, isParent in NXSL class "NetObj"
- New NXSL function GetCurrentTimeMs
- Fixed possible server crash when agent tunnel is closing
- Always use UTF-8 instead of current system character encoding for exported CSV files
- Fixed process handle leak in Windows agent when external parameter providers are used
- Westerstrand SNMP device driver can read device geolocation
- NXSL function PollerTrace works in autobind filter scripts and instance discovery filter scripts
- Improved diagnostic output for manually started polls
- Improved network map tooltip data update

Best regards,
Victor

For zoned environments probably the only way is to use filtering script that will check allowed ranges per zone. Mapping tables or custom attributes on zone objects can be used to store allowed or forbidden ranges.
But in general I think we need per zone address range filters, probably implemented as zone object properties.

Best regards,
Victor

Most likely server processes already queued discovery tasks. Restart of netxmsd should fix it.

Best regards,
Victor

Currently it's not possible. I just added new function for that - GetCurrentTimeMs(). It will be available in next patch release.

Best regards,
Victor

Quote from: Akira on March 02, 2021, 08:08:25 AM

Quote from: Victor Kirhenshtein on March 01, 2021, 06:55:21 PM
Ну значит он падает, а не виснет. Сможете прислать дампы?

сервис запущен был.
Там дамп 200мб, как передать?

Можно сюда залить: https://cloud.radensolutions.com/s/GAWHMA9ppyz78LF

Hi,

second issue (404 errors for certain parameters) are normal system activity - server requests various data from nodes, but not every node support all parameters, so 404 response is expected.

PDH error is a bit more strange. When collecting performance counter metadata, agent reads all available objects and counters for each object. Such error means that PDH API reports object, but on attempt to read counters for that object returns error that such object is not found. We will add actual object name to this error message in next patch release - it will help to debug this issue further.

Best regards,
Victor

Yes, it is working. Make sure that nxmc really picks up Java 11.

Best regards,
Victor

File nxmc.properties should be placed anywhere in class path. For Tomcat 9 good choice will be shared class path, by default it is $CATALINA_BASE/lib (unless redefined by the common.loader property in $CATALINA_BASE/conf/catalina.properties). You can find more details on Tomcat 9 class loader here: https://tomcat.apache.org/tomcat-9.0-doc/class-loader-howto.html.

Best regards,
Victor

Ну значит он падает, а не виснет. Сможете прислать дампы?

Rule number 5 - it has no condition, so it catches every event, and sends notification.

Best regards,
Victor

Most likely you have some sort of "catch all" rule that also triggered for those events without explicitly listing them (like "all events from object"). If you can show your full policy I can comment further.

Best regards,
Victor

Hi,

seems that there are two different problems. One is with CSV export - currently management console uses current system encoding instead of UTF-8, which can produce files like you have shown. I just fixed that, next version will always use UTF-8 for CSV encoding.
Second issue is likely database encoding - as we are using varchar fields text is converted to whatever character set is choosen for database, with all unsupported Unicode characters converted to ?. Easiest option probably would be to switch database to UTF-8 encoding, but is is only available for SQL Server 2019. If you only encounter such characters in software inventory, you can try to change data type for columns in table software_inventory to NVARCHAR from VARCHAR. I have registered issue in our bugtracker for considering switching to NVARCHAR for MS SQL: https://track.radensolutions.com/issue/NX-2016 - if implemented it will solve issues like this one.

Best regards,
Victor

Hi,

I would create script DCI that will read last known values of two other DCIs and return two distinctive values for match and not match. Then set threshold on it as usual.

Best regards,
Victor

Hi,

if you got emails that means that somewhere you have rule that catches this event and sends an email. Maybe something  generic, like all events from certain devices or with certain severity. Easiest way to turn off processing for this event is to add as first rule in EPP with condition set to just that event and action set to "stop processing".

Best regards,
Victor

Hi,

currently the only way is to setup event processing action for desired events that will execute external script that will use snmptrap command to send trap to upper level NMS. We plan to have SNMP trap sending as separate action type for some time already, it will be implemented eventually. 

Best regards,
Victor