Details
-
Type:
Enhancement
-
Status:
Resolved
-
Priority:
Major
-
Resolution: Fixed
-
Affects Version/s: 1.8.12
-
Component/s: Provisioning/Discovery/Capsd/Provisiond
-
Security Level: Default (Default Security Scheme)
-
Labels:None
Description
I have started using the provisioner to provision nodes automatically, and quite often I run into issues where the nodes will never be pushed to the database. This happens for example when identical nodes are provisioned, causing the import process to halt right after reaching the duplicate node.
On my current setup, I have some 300 nodes still waiting to be provisioned for some unknown reason, yet provisiond does not complain about anything, nor can I find anything else in any log files.
The import process should be somewhat verbosely logged, in order for the administrator to easily find the problem.
This is an enhancement request, requesting better logging under general production, to make handling of errors during database imports through provisiond better.
On my current setup, I have some 300 nodes still waiting to be provisioned for some unknown reason, yet provisiond does not complain about anything, nor can I find anything else in any log files.
The import process should be somewhat verbosely logged, in order for the administrator to easily find the problem.
This is an enhancement request, requesting better logging under general production, to make handling of errors during database imports through provisiond better.
Digging around the code, there is actually a decent amount of debugging in the code, it's just only enabled for DEBUG level and higher (which is not the default in stable releases).
You can get better feedback by bumping logging on Provisiond up to DEBUG in $OPENNMS_HOME/etc/log4j.properties
That said, the logging needed a bit for cleanup still, we should change the logging to default to INFO level, at least, and any time we start a new scan of a resource, log that at INFO. That I've done in the 1.10 and 1.11 trees.