Provisiond accepts multiple primary SNMP interfaces
Description
Acceptance / Success Criteria
None
Lucidchart Diagrams
Activity
Show:
Chandra Gorantla May 13, 2020 at 1:26 PM
UI changes are added with commit 53736119a4d0a3ad6700deb996e353afc8fe9545
Patrick Schweizer May 10, 2020 at 11:45 PM
The frontend still needs to generate a sensible message.
@Chandra Gorantla, thanks for taking over!
Patrick Schweizer April 10, 2020 at 7:38 PM
Patrick Schweizer April 7, 2020 at 4:16 PM
Goal: Don't accept input with 2 primary interfaces.
Input can come from
Rest API: reject and give indication over API
Scheduler
Marcel Fuhrmann March 18, 2020 at 5:45 PM
The only way right now to avoid it is to create correct provisioning requisitions. In my case the requisitions came from PRIS (XLS source). Someone defined the node wrong.
But this can also happen in the Provisiond requisition UI. If you add a second interface SNMP Primary flag is set to N by default, but you can set P manually and save the requisition.
Fixed
Details
Assignee
Chandra GorantlaChandra GorantlaReporter
Marcel FuhrmannMarcel FuhrmannLabels
Components
Sprint
NoneFix versions
Affects versions
Priority
Major
Details
Details
Assignee
Chandra Gorantla
Chandra GorantlaReporter
Marcel Fuhrmann
Marcel FuhrmannLabels
Components
Sprint
None
Fix versions
Affects versions
Priority
PagerDuty
PagerDuty Incident
PagerDuty
PagerDuty Incident
PagerDuty

PagerDuty Incident
Created March 17, 2020 at 6:19 PM
Updated May 19, 2020 at 4:59 PM
Resolved May 19, 2020 at 4:59 PM
Provisiond accepts node definitions with two primary tagged snmp interfaces.
<node node-label="test" foreign-id="test"> <interface ip-addr="192.168.0.10" snmp-primary="P"> <monitored-service service-name="ICMP"/> </interface> <interface ip-addr="62.209.55.22" snmp-primary="P"> <monitored-service service-name="ICMP"/> <monitored-service service-name="SNMP"/> </interface> </node>
The result is, that OpenNMS does not collect any interface table metrics from this device without any hint why (or maybe I didn't find one).
We really have too many places why a collection can fail. See here: https://opennms.discourse.group/t/troubleshoot-performance-data-collection-from-snmp-agents/
It was not obvious to search in provisiond for a collection issue.
I think provisiond should complain about not valid node definitions and raise an alarm, so a administrator is able to notice it somehow.
But it should import all other nodes with correct definitions in that case.