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 51027: Track allocation candidates to bound allocator.
Date Thu, 08 Sep 2016 04:36:15 GMT


> On Sept. 3, 2016, 9:21 a.m., Guangya Liu wrote:
> > src/master/allocator/mesos/hierarchical.cpp, lines 484-485
> > <https://reviews.apache.org/r/51027/diff/4/?file=1490734#file1490734line484>
> >
> >     I did some test with benchmark test of `SlaveAndFrameworkCount/HierarchicalAllocator_BENCHMARK_Test.AddAndUpdateSlave/0`
and it failed as following:
> >     
> >     ```
> >     ../../src/tests/hierarchical_allocator_tests.cpp:3459: Failure
> >     Value of: offerCallbacks.load()
> >       Actual: 5
> >     Expected: slaveCount
> >     Which is: 1000
> >     [  FAILED  ] SlaveAndFrameworkCount/HierarchicalAllocator_BENCHMARK_Test.AddAndUpdateSlave/0,
where GetParam() = (1000, 1) (497 ms)
> >     [----------] 1 test from SlaveAndFrameworkCount/HierarchicalAllocator_BENCHMARK_Test
(498 ms total)
> >     
> >     [----------] Global test environment tear-down
> >     [==========] 1 test from 1 test case ran. (512 ms total)
> >     [  PASSED  ] 0 tests.
> >     [  FAILED  ] 1 test, listed below:
> >     [  FAILED  ] SlaveAndFrameworkCount/HierarchicalAllocator_BENCHMARK_Test.AddAndUpdateSlave/0,
where GetParam() = (1000, 1)
> >     
> >      1 FAILED TEST
> >     ```
> >     
> >     The reason is that we cannot make sure all of the `_allocate()` finished after
`addSlave` finished.
> >     
> >     Shall we do a `while` loop in the benchmark to wait till all allocations are
got?
> 
> Guangya Liu wrote:
>     Perhaps we can simply remove the `ASSERT_EQ` for offer and agent count in `HierarchicalAllocator_BENCHMARK_Test.AddAndUpdateSlave`
as this patch will not trigger `allocate()` synchronously, so we cannot make sure the agent
count is always same as offer count.

Good catch, I'll run the benchmarks and update https://reviews.apache.org/r/51028/


- Jacob


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/51027/#review147722
-----------------------------------------------------------


On Sept. 3, 2016, 6:05 a.m., Jacob Janco wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/51027/
> -----------------------------------------------------------
> 
> (Updated Sept. 3, 2016, 6:05 a.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.
> - Batched allocations are handled synchronously.
> 
> 
> Diffs
> -----
> 
>   src/master/allocator/mesos/hierarchical.hpp dd07ed221d2c1755d2478369641ffdc46ecc4471

>   src/master/allocator/mesos/hierarchical.cpp 9e5db2196c6a541dc1208ba8b9f13ef9a518bcc4

> 
> Diff: https://reviews.apache.org/r/51027/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> note: check without filters depends on https://reviews.apache.org/r/51028
> 
> With new benchmark https://reviews.apache.org/r/49617: 
> Sample output without 51027:
> [ RUN      ] SlaveAndFrameworkCount/HierarchicalAllocator_BENCHMARK_Test.FrameworkFailover/22
> Using 10000 agents and 3000 frameworks
> Added 3000 frameworks in 57251us
> Added 10000 agents in 3.21345353333333mins
> allocator settled after  1.61236038333333mins
> [       OK ] SlaveAndFrameworkCount/HierarchicalAllocator_BENCHMARK_Test.FrameworkFailover/22
(290578 ms)
> 
> Sample output with 51027:
> [ RUN      ] SlaveAndFrameworkCount/HierarchicalAllocator_BENCHMARK_Test.FrameworkFailover/22
> Using 10000 agents and 3000 frameworks
> Added 3000 frameworks in 39817us
> Added 10000 agents in 3.22860541666667mins
> allocator settled after  25.525654secs
> [       OK ] SlaveAndFrameworkCount/HierarchicalAllocator_BENCHMARK_Test.FrameworkFailover/22
(220137 ms)
> 
> 
> Thanks,
> 
> Jacob Janco
> 
>


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