NetXMS Support Forum

Hi all!

NetXMS version 1.2.17 is out. Changes since previous release:

- New DCI source: NXSL script executed on server
- Configurable node matching policy for built-in syslog server (controlled by SyslogNodeMatchingPolicy configuration parameter)
- Oracle monitoring subagent improved (bugs fixed, new metrics)
- nxalarm tool supports commands add-comment and get-comments
- Source port number added to events generated from SNMP trap (available via "sourcePort" named parameter)
- Jira link: added possibility to set project's component for issues being created
- ICMP proxy can be set for nodes
- New methods in NXSL classes Node, Interface, and NetObj: setStatusCalculation and setStatusPropagation
- New attributes "slot" and "port" in NXSL class Interface
- Can execute arbitrary NXSL script in context of node, cluster, subnet, or container object from management console
- Improved network topology changes detection
- Added driver for H3C switches
- Management console:
    - Can show alarms for multiple selected objects
    - Fixed non-working ordering in event list in alarm details view
    - Fixed bug with LDAP user system rights
    - Added default search string for LDAP to select all objects: "(objectClass=*)"
    - Added device geolocation tracking and display on map
    - Filter in event processing policy editor
    - Fixed bug with deletion of subnet with corrected IP
    - Added command line option to open specific dashboard after login
- Android console:
   - Fixed issue #661
   - Added node boot time to overview tab
   - Fix bug in notifying connection point not found
   - Fix bug in computing interface expanded list size (removed hardcoded values)
   - Added manage, unmanage, set expected state (up, down, ignore) and find switch port to interface list
   - Fix aesthetic problems in expandable list (graphs and interfaces list)
   - Integration of new support library
   - Target to new API version (20)
- Fixed issues: #51, #91, #324, #359, #409, #532, #542, #571, #563, #590, #593, #613, #617, #624, #626, #629, #632, #635, #638, #639, #641, #642, #649, #658, #663

Best regards,
Victor

That's the historical problem - event parameters for recovery event goes in different order than for activation event. If you want to use single policy you should access them by names instead of positions, like %<dciDescription> instead of %2. You can use the following names: dciId, dciName, dciDescription, thresholdValue, currentValue, instance, isRepeatedEvent.

Best regards,
Victor

yes, I've made changes in topology discovery module. Now it should correctly remove outdated information (it does on my tests).

Best regards,
Victor

Hi,

so you didn't get WARNING message when disk space increased let's say from 4% to 7%? That's strange because your setup looks correct and you should get WARNING event when threshold "< 5" deactivates. How did you check that you got messages?

Best regards,
Victor

Hi,

I've added configuration option EscapeLocalCommands. If it is set to 1, tab, carriage return, and new line characters will be replaced by \t, \r, and \n in commands. It will be available in release 1.2.17.

Best regards,
Victor

3 секунды.

Yes, thank you. There was a bug (related to MS SQL only), I've fixed it. Release 1.2.17 will contain this fix (we plan to release it tomorrow).

Best regards,
Victor

Hi,

can you please send definition of table "items" as it is created in your database?

Best regards,
Victor

Hi,

currently it's not possible. We can add appropriate hooks in the future, as idea looks useful. I'll regsiter this as feature request.

Best regards,
Victor

В настройках субагента portcheck можно задать таймаут параметром

Timeout = milliseconds

это будет значение по умолчанию для всех проверок. Кроме того таймаут можно задать прямо в запросе последним аргументом, например так:

ServiceCheck.HTTP(10.0.0.1,80,/,,"^HTTP/1\.[01]",2000)

(таймаут 2 секунды).

Вообще ServiceCheck.HTTP имеет 6 аргументов: хост, порт, URI, дополнительный заголовок, regexp для валидации, таймаут.

А информация по трафику собирается через SNMP или с агента? Если с агента, то все паранетры Net.Interface. позволяют указывать имя интерфейса вместо индекса.

Да, был баг что статус не пересчитывался. Исправил.

Hi,

#2 is implemented in 1.2.17. This is turned off by default to maintain compatibility with previous versions. Setting parameter ProcessOfflineEvents in LogWatch section to yes or true will enable this feature. When on, agent will save timestamp of last processed event and on start will process all events since last processed.

Best regards,
Victor

Hi,

try to set debug level to at least 4 (add DebugLevel = 4 to netxmsd.conf) and check server's log. If you have logging set to event log, I suggest to change it to file first.

Best regards,
Victor

Hi,

IP topology map shows IP topology ignoring physical connectivity between nodes - and I suppose that all nodes and switches in your network are in the same IP subnet. What you need is layer 2 topology map, which shows physical connections between devices.

Best regards,
Victor