Scheduler Failures are no longer logged with enough detail for a site admin to do problem determination
Bug #1431291 reported by
Sean Dague
This bug affects 5 people
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Invalid
|
High
|
Pranav Salunke |
Bug Description
After the move of the scheduler dispatch into conductor all the logging changed. The impact is that the scheduler no longer logs scheduled failures at WARN in the scheduler log with the instance id in question. (This apparently also somehow slipped back into Juno)
This is a notable regression from an ops perspective, and has required everyone to start running debug and doing lots of log searching to figure out whats going on.
The ops community asked that we log at a high level (WARN seems appropriate here) with reason, context, and instance id which makes their problem determination loop on "nova didn't boot my thing" from users much quicker.
Changed in nova: | |
status: | New → Confirmed |
importance: | Undecided → High |
milestone: | none → kilo-3 |
Changed in nova: | |
assignee: | nobody → Sylvain Bauza (sylvain-bauza) |
tags: | added: scheduler |
Changed in nova: | |
milestone: | kilo-3 → kilo-rc1 |
tags: | removed: kilo-rc-potential |
Changed in nova: | |
status: | In Progress → Confirmed |
assignee: | Sylvain Bauza (sylvain-bauza) → nobody |
Changed in nova: | |
assignee: | nobody → Pranav Salunke (dguitarbite) |
To post a comment you must log in.
Related fix proposed to branch: master /review. openstack. org/163771
Review: https:/