-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13621/#review25245
-----------------------------------------------------------

Ship it!



docs/Upgrades.md
<https://reviews.apache.org/r/13621/#comment49542>

    I wouldn't say in favor of /health. Removing /vars and adding /health are 
orthogonal, but happen to both be present in 0.14.0. So maybe just mention that 
/vars no longer exists?



docs/Upgrades.md
<https://reviews.apache.org/r/13621/#comment49546>

    Ditto, we should mention this in the changelog summary as well.



docs/Upgrades.md
<https://reviews.apache.org/r/13621/#comment49547>

    Upgrade the scheduler to be setting FrameworkInfo.checkpoint, is this 
mandatory? Shouldn't we say if checkpointing is desired?



docs/Upgrades.md
<https://reviews.apache.org/r/13621/#comment49548>

    It might not be totally clear what "upgrade" means here, install the new 
binaries without restarting the slaves?
    
    e.g.
    Upgrade the slaves (install the new slave binaries).
    
    Would be great to cc Brenden on this or someone else who runs a cluster, to 
see how clear these instructions are. I'm biased so it seems intuitive for me 
but hard to tell how this may read to someone less familiar with upgrades. :)


- Ben Mahler


On Aug. 16, 2013, 6:02 p.m., Vinod Kone wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/13621/
> -----------------------------------------------------------
> 
> (Updated Aug. 16, 2013, 6:02 p.m.)
> 
> 
> Review request for mesos, Benjamin Hindman, Ben Mahler, and Thomas Marshall.
> 
> 
> Repository: mesos-git
> 
> 
> Description
> -------
> 
> Also cherrypicked 0.13.0 upgrade instructions.
> 
> 
> Diffs
> -----
> 
>   docs/Upgrades.md 71054828f1269f8eb7dafd495204ff3c154bc409 
> 
> Diff: https://reviews.apache.org/r/13621/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Vinod Kone
> 
>

Reply via email to