Skip to content
This repository has been archived by the owner on Oct 23, 2024. It is now read-only.

Killing a lost resident task results in expunge #5207

Closed
timcharper opened this issue Feb 18, 2017 · 3 comments
Closed

Killing a lost resident task results in expunge #5207

timcharper opened this issue Feb 18, 2017 · 3 comments
Assignees
@timcharper

Comments

@timcharper
Copy link
Contributor

timcharper commented Feb 18, 2017 •
edited by marcomonaco
Loading

In KillServiceActor.scala we decide to expunge a task if it is lost or terminated. This should be fixed such that we transition the task back to Reserved, and only expunge if explicitly asked to do so.
@timcharper timcharper self-assigned this Feb 18, 2017
@timcharper timcharper added this to the Marathon 1.4 milestone Feb 18, 2017
@timcharper timcharper added the bug label Feb 18, 2017
@timcharper timcharper mentioned this issue Feb 18, 2017
Closed
6 tasks
@meichstedt
Copy link
Contributor

@timcharper I see you have this in progress but since it's a public holiday on your side and I had some time while being on-call, I've provided https://phabricator.mesosphere.com/D529

Please consider this a WIP and feel free to take over if you see fit.

@timcharper
Copy link
Contributor Author

Thanks @meichstedt for starting this ! I'll commandeer the change and push it to completion.

@timcharper
Copy link
Contributor Author

I've finished the patch and it's confirmed to work

@timcharper timcharper mentioned this issue Feb 22, 2017
Closed
2 tasks
timcharper pushed a commit that referenced this issue Feb 22, 2017
@meichstedt
Refine logic on how to treat instances that shall be killed.
ef6bd77 
Summary:
This will transition unreachable or reachable stateful instances to Reserved when they're about to be killed.

Fixes #5207

Also-By: tharper@mesosphere.com

Test Plan: sbt test

Reviewers: unterstein, jasongilanfarr, jenkins, meichstedt

Reviewed By: jenkins

Subscribers: jdef, marathon-team

Differential Revision: https://phabricator.mesosphere.com/D529
timcharper pushed a commit that referenced this issue Feb 22, 2017
@meichstedt
Refine logic on how to treat instances that shall be killed.
eeed52b 
Summary:
This will transition unreachable or reachable stateful instances to Reserved when they're about to be killed.

Fixes #5207

Also-By: tharper@mesosphere.com

Test Plan: sbt test

Reviewers: unterstein, jasongilanfarr, jenkins, meichstedt

Reviewed By: jenkins

Subscribers: jdef, marathon-team

Differential Revision: https://phabricator.mesosphere.com/D529
@timcharper timcharper mentioned this issue Feb 28, 2017
Closed
timcharper pushed a commit that referenced this issue Mar 6, 2017
@meichstedt
Don't destroy persistent volumes when killing unreachable resident tasks
fec2729 
Summary:
Killing an unreachable resident task will do nothing, rather than
destroy the reservations.

Fixes #5207

Also-By: tharper@mesosphere.com

Test Plan: sbt test

Reviewers: unterstein, meichstedt, jasongilanfarr, jenkins

Reviewed By: meichstedt, jasongilanfarr, jenkins

Subscribers: jdef, marathon-team

Differential Revision: https://phabricator.mesosphere.com/D529
@timcharper timcharper mentioned this issue Mar 6, 2017
Merged
@d2iq-archive d2iq-archive locked and limited conversation to collaborators Mar 27, 2017
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Assignees

@timcharper timcharper

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@timcharper @meichstedt @marcomonaco