mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jacob Janco <jjanco....@gmail.com>
Subject Re: Review Request 53479: Perform agent GC asynchronously.
Date Tue, 18 Jul 2017 22:15:48 GMT


> On July 17, 2017, 8:43 p.m., Jiang Yan Xu wrote:
> > src/slave/gc.hpp
> > Lines 114-117 (original), 115-118 (patched)
> > <https://reviews.apache.org/r/53479/diff/8/?file=1777247#file1777247line116>
> >
> >     I see that `promise == that.promise` here is required by the Multimap but does
it make sense to have `PathInfo` equality depend only on the path?
> >     
> >     I don't think `Promise` is supposed to be comparable.
> 
> Jiang Yan Xu wrote:
>     OK I think strictly speaking we shouldn't assume that there are no equal values under
the same key in a Multimap (or std::multimap).
>     
>     This case shouldn't happen with our GC so we should probably insert some hard `CHECK`s
for it (in a separate patch) but I don't think removing `promise == that.promise` has anything
to do with it, i.e., the problem exists today.
> 
> Jacob Janco wrote:
>     Updating and removing the blocking dependency. Actually, I don't think we even use
this operator...why not just remove this and add the checks like you said.
> 
> Jiang Yan Xu wrote:
>     I think it's required by Multimap.

Per discussion: Now that we store a pointer remove this equality.


- Jacob


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/53479/#review180730
-----------------------------------------------------------


On July 18, 2017, 12:23 a.m., Jacob Janco wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/53479/
> -----------------------------------------------------------
> 
> (Updated July 18, 2017, 12:23 a.m.)
> 
> 
> Review request for mesos, Benjamin Mahler and Jiang Yan Xu.
> 
> 
> Bugs: MESOS-6549
>     https://issues.apache.org/jira/browse/MESOS-6549
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> - Previously, rmdir operations were serialized
>   through the garbage collector. Expensive removal
>   events had the potential to delay task launch.
> - MESOS-6549
> 
> 
> Diffs
> -----
> 
>   src/slave/gc.hpp 5ea82456cffa374203371f227b6f0ee00296553e 
>   src/slave/gc.cpp 961f547793984449ea113d9664b12b5033638c58 
> 
> 
> Diff: https://reviews.apache.org/r/53479/diff/9/
> 
> 
> Testing
> -------
> 
> `./bin/mesos-tests.sh --gtest_filter="GarbageCollector*" --gtest_repeat=100 --gtest_break_on_failure`
> 
> 
> Thanks,
> 
> Jacob Janco
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message