mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Mann <g...@mesosphere.io>
Subject Re: Review Request 70702: Sequentialized all events to master's `/api/v1` subscribers.
Date Wed, 12 Jun 2019 13:31:31 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/70702/#review215838
-----------------------------------------------------------


Ship it!





src/master/master.cpp
Lines 12933-12937 (patched)
<https://reviews.apache.org/r/70702/#comment302774>

    Unfortunately we don't have a benchmark which would allow us to assess the performance
impact of the additional copying here. Since the objects are small, I'm hopeful that it isn't
significant.


- Greg Mann


On May 23, 2019, 4:17 a.m., Chun-Hung Hsiao wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/70702/
> -----------------------------------------------------------
> 
> (Updated May 23, 2019, 4:17 a.m.)
> 
> 
> Review request for mesos, Benjamin Mahler and Greg Mann.
> 
> 
> Bugs: MESOS-9785
>     https://issues.apache.org/jira/browse/MESOS-9785
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> The master needs to create object approvers before sending an event to
> its `/api/v1` subscribers. The creation calls `process::collect`, which
> does not have any ordering guarantee. As a result, events might be
> reordered, which could be unexpected by subscribers.
> 
> This patch imposes an order between events by sequentializing the
> creation of object approvers. The actual creations can still go in
> parallel, but the returned futures will be completed in the creation
> order.
> 
> 
> Diffs
> -----
> 
>   src/master/master.hpp 2771f4c045c877b7d8aa5db042810232c0e40ba0 
>   src/master/master.cpp 4d7c37cf8814e45432b3fe15173f5343676a372b 
> 
> 
> Diff: https://reviews.apache.org/r/70702/diff/2/
> 
> 
> Testing
> -------
> 
> make check
> 
> Ran the test updated in r/70651 for 500 times under stress.
> 
> 
> Thanks,
> 
> Chun-Hung Hsiao
> 
>


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