-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/71476/#review217826
-----------------------------------------------------------


Fix it, then Ship it!




Oh, that's a long-awaited fix!

Looks good to me, albeit there is a corner case around which I cannot wrap my 
head (see below).


src/master/allocator/mesos/hierarchical.cpp
Lines 1489-1494 (original), 1523-1528 (patched)
<https://reviews.apache.org/r/71476/#comment305220>

    Previously, this code was executed even in case of recovering resources of 
already removed slave.
    
    Now I don't see any framework untracking code in `removeSlave()` or 
`untrackAllocatedResources()`. 
    
    Does this mean that now we are potentially leaking frameworks? Or not?


- Andrei Sekretenko


On Sept. 12, 2019, 10:42 p.m., Meng Zhu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/71476/
> -----------------------------------------------------------
> 
> (Updated Sept. 12, 2019, 10:42 p.m.)
> 
> 
> Review request for mesos, Andrei Sekretenko and Benjamin Mahler.
> 
> 
> Bugs: MESOS-621
>     https://issues.apache.org/jira/browse/MESOS-621
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Simplified recover resources when removing frameworks or agents.
> 
> 
> Diffs
> -----
> 
>   src/master/allocator/mesos/hierarchical.hpp 
> 5ea37912dadded47eaa2d354889c95533b191c59 
>   src/master/allocator/mesos/hierarchical.cpp 
> 256fdce61ccfb768605cac1579c9a6632cd26fec 
>   src/master/master.cpp a2c289a8b6e30eac690f77c7fc1b12b16f62f541 
> 
> 
> Diff: https://reviews.apache.org/r/71476/diff/1/
> 
> 
> Testing
> -------
> 
> make check
> New test in https://reviews.apache.org/r/71479/
> 
> 
> Thanks,
> 
> Meng Zhu
> 
>

Reply via email to