Failed to recover paused instance

Bug #1782516 reported by Rikimaru Honjo
12
This bug affects 2 people
Affects Status Importance Assigned to Milestone
masakari
In Progress
Undecided
takahara.kengo

Bug Description

[Error]
Recovering host-failure was failed when there was paused state instance on the failed host.
As a result, notification status became "failed".
(Instance's state became "ACTIVE".)

And, the following message was outputted in masakari's log.

Cannot 'stop' instance <UUID> while it is in task_state rebuilding

I used the latest version of masakari.

[Cause of error]
Masakari will try to do the following processes for a paused instance when host-failure is occurred.

1. Call reset state API for the instance.
-> The instance's vm_state will be error. And task_state will be none.
2. Call evacuate API for the instance.
3. Wait until completing evacuation.
4. Call stop API for the instance.

In my opinion, there is a bug in step 3.
It will immediately go to step 4 if vm_state is error.
As a result, stop API will be failed.
Because task_state=rebuilding in that time.

description: updated
Changed in masakari:
assignee: nobody → takahara.kengo (takahara.kengo)
Revision history for this message
takahara.kengo (takahara.kengo) wrote :
Changed in masakari:
status: New → In Progress
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.