Race in maas setup/teardown

Bug #1658837 reported by Curtis Hovey
4
This bug affects 1 person
Affects Status Importance Assigned to Milestone
juju-ci-tools
Triaged
Low
Unassigned

Bug Description

As seen at
   http://reports.vapour.ws/releases/issue/57c721c7749a5615669b72dc
   http://reports.vapour.ws/releases/issue/56f3c6ae749a5672accf4402

The endpoint-bindings test does not wait for machines and resources to enter the right state before adding/removing subnets and spaces. Clean up scripts were written to recover from the problem, but they too can fail because of race conditions.

The mass cleanup script that resets networks in the bindings test failed to match the hostname (machine=) to look up the system id that needs reseting. the test failed to setup.

++ grep 10.0.40.0/24@
++ cut -d @ -f2
+ system_id=
+ maas env20 machine release

^ We expected the grep to be for name like juju-qa-maas-node-40 but
machine=10.0.40.0/24
was selected previously

Revision history for this message
Curtis Hovey (sinzui) wrote :

This issue is not high if the logout race if fixed.

Changed in juju-ci-tools:
importance: High → Medium
Curtis Hovey (sinzui)
summary: - CI maas cleanup script selected network instead of hostname
+ Race in maas setup/teardown
description: updated
description: updated
Curtis Hovey (sinzui)
Changed in juju-ci-tools:
importance: Medium → Low
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.