mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Guangya Liu <gyliu...@gmail.com>
Subject Re: Review Request 51027: Track allocation candidates to bound allocator.
Date Mon, 17 Oct 2016 05:34:20 GMT


> On 十月 4, 2016, 9:47 p.m., Guangya Liu wrote:
> > Jacob, two other comments for this:
> > 
> > 1) when will you have the patch `SmallOfferFilter` ready for review?
> > 2) I think we may need another benchmark test which only do two operations: a) add
multiple agents first 2) add multiple frameworks, this test can make sure we `batched` the
`allocate()` request for all frameworks.
> 
> Jacob Janco wrote:
>     Added the fix for you: https://reviews.apache.org/r/52534

Are you planning to add a new benchmark test of add agents first then add frameworks to test
your patch?


> On 十月 4, 2016, 9:47 p.m., Guangya Liu wrote:
> > src/master/allocator/mesos/hierarchical.cpp, line 1306
> > <https://reviews.apache.org/r/51027/diff/9/?file=1522284#file1522284line1306>
> >
> >     The `addFramework` will call `allocate()` and allocate `all agents resources`
but not a `single agnent`, so here we should not mention `addFramework`, but only the following
three:
> >     
> >     1) addSlave
> >     2) updateSlave
> >     3) updateUnavailability?
> 
> Jacob Janco wrote:
>     I mentioend addFramework and addSlave intentionally so we have an example of allocating
all known slaves as well as for a single slave. In both cases we will take the union of allocationCandidates.
For example I addSlave(agent1) and immediately addFramework() ... an allocation would be queued
for the single agent then all agents would be added in the union when addFramework() is called.
When the actual allocation run occurs all slaves known at the time of addFramework() will
be included in allocationCandidates.

That make sense, thanks Jacob.


- Guangya


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/51027/#review151408
-----------------------------------------------------------


On 十月 4, 2016, 11:31 p.m., Jacob Janco wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/51027/
> -----------------------------------------------------------
> 
> (Updated 十月 4, 2016, 11:31 p.m.)
> 
> 
> Review request for mesos, Benjamin Mahler, Guangya Liu, James Peach, Klaus Ma, and Jiang
Yan Xu.
> 
> 
> Bugs: MESOS-3157
>     https://issues.apache.org/jira/browse/MESOS-3157
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> - Triggered allocations dispatch allocate() only
>   if there is no pending allocation in the queue.
> - Allocation candidates are accumulated and only
>   cleared when enqueued allocations are processed.
> 
> 
> Diffs
> -----
> 
>   src/master/allocator/mesos/hierarchical.hpp 2c31471ee0f5d6836393bf87ff9ecfd8df835013

>   src/master/allocator/mesos/hierarchical.cpp c8f9492ee1b69e125a1e841116d22a578a9b524e

> 
> Diff: https://reviews.apache.org/r/51027/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> note: check without filters depends on https://reviews.apache.org/r/51028 and https://reviews.apache.org/r/52534
> 
> 
> Thanks,
> 
> Jacob Janco
> 
>


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