> On April 29, 2015, 9:50 p.m., Kevin Sweeney wrote:
> > Actually I may have jumped the gun here - can you verify that mesos will 
> > behave sanely when we flip frameworkInfo.checkpoint from false to true (as 
> > everyone who has been using the defaults in 0.7.0 will encounter when they 
> > attempt to upgrade to 0.8.0)?

I observed sane behavior (at least, a functioning cluster) with the following 
sequence:
- create a new cluster in vagrant on master
- curl :5051/master/state.json, confirm checkpoint=false
- apply this patch, rebuild/restart scheduler
- curl :5051/master/state.json, observed checkpoint=false (despite sending 
checkpoint=true)
- end-to-end tests pass
- restart master, checkpoint=true is picked up


- Bill


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/33676/#review82040
-----------------------------------------------------------


On April 29, 2015, 3:04 p.m., Bill Farner wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/33676/
> -----------------------------------------------------------
> 
> (Updated April 29, 2015, 3:04 p.m.)
> 
> 
> Review request for Aurora and Kevin Sweeney.
> 
> 
> Bugs: AURORA-716
>     https://issues.apache.org/jira/browse/AURORA-716
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Always require slave checkpointing.
> 
> 
> Diffs
> -----
> 
>   
> src/main/java/org/apache/aurora/scheduler/mesos/CommandLineDriverSettingsModule.java
>  d0994203b5650f44ca2eb32e1e2aa61875163854 
> 
> Diff: https://reviews.apache.org/r/33676/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Bill Farner
> 
>

Reply via email to