-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/69588/#review212376
-----------------------------------------------------------



It looks like this patch does clean up semantics around unset role values. 
Looking at e.g., the authz for `UNRESERVE`, when previously a resource was 
unreserved (neither `reservations` nor `role`), we would authorize a reserve 
against an empty role. I am not sure we change behavior observable to 
authorizers here. If we do, that's fine and makes sense, but should probably be 
called out.

- Benjamin Bannier


On Dec. 20, 2018, 12:20 a.m., Chun-Hung Hsiao wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/69588/
> -----------------------------------------------------------
> 
> (Updated Dec. 20, 2018, 12:20 a.m.)
> 
> 
> Review request for mesos, Benjamin Bannier, Benjamin Mahler, Michael Park, 
> and Till Toenshoff.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Since the master now upgrades resources to the post-reservation-refinement
> format before authorization (see MESOS-7735), the authorization logic in the
> master becomes outdated. This patch cleans it up.
> 
> 
> Diffs
> -----
> 
>   src/master/master.cpp b4faf2b077a0288ba36195b7a21402932489d316 
> 
> 
> Diff: https://reviews.apache.org/r/69588/diff/2/
> 
> 
> Testing
> -------
> 
> sudo make check
> 
> 
> Thanks,
> 
> Chun-Hung Hsiao
> 
>

Reply via email to