mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Benno Evers <bev...@mesosphere.com>
Subject Re: Review Request 63368: Added MemoryProfiler class to Libprocess.
Date Mon, 16 Apr 2018 14:55:28 GMT


> On April 16, 2018, 10:32 a.m., Alexander Rukletsov wrote:
> > 3rdparty/libprocess/src/memory_profiler.cpp
> > Lines 286 (patched)
> > <https://reviews.apache.org/r/63368/diff/10/?file=1998458#file1998458line286>
> >
> >     Can we add a subfolder with current process pid to address scenarios where multiple
libprocess processes are profiled on the same machine?

The `XXXXXX`-part already makes it unique per process.


> On April 16, 2018, 10:32 a.m., Alexander Rukletsov wrote:
> > 3rdparty/libprocess/src/memory_profiler.cpp
> > Lines 325-328 (patched)
> > <https://reviews.apache.org/r/63368/diff/10/?file=1998458#file1998458line325>
> >
> >     This can also happen if there are no useful data in the profile. Is it what
you mean by "empty"?

Yes, that's what I meant. Reworded to "contains data".


> On April 16, 2018, 10:32 a.m., Alexander Rukletsov wrote:
> > 3rdparty/libprocess/src/memory_profiler.cpp
> > Lines 667-672 (patched)
> > <https://reviews.apache.org/r/63368/diff/10/?file=1998458#file1998458line667>
> >
> >     Current logic allows users to download profiles from the _previous_ run while
the _current_ run is active. Is it done on purpose? 
> >     
> >     I can imagine users expecting download to return intermediate results of the
_current_ run, hence I'd rather either clean the state before starting a new run or disallow
implicit, i.e., without id, download request if a profiling is running.

Ok, added as a special case.


- Benno


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63368/#review201184
-----------------------------------------------------------


On April 10, 2018, 2:39 p.m., Benno Evers wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/63368/
> -----------------------------------------------------------
> 
> (Updated April 10, 2018, 2:39 p.m.)
> 
> 
> Review request for mesos, Alexander Rukletsov and Benjamin Mahler.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This class exposes profiling functionality of jemalloc memory allocator
> when it is detected to be the memory allocator of the current process.
> 
> In particular, it gives developers an easy method to collect and
> access heap profiles which report which pieces of code were
> responsible for allocating memory.
> 
> 
> Diffs
> -----
> 
>   3rdparty/libprocess/Makefile.am c895d3ac7b9cc5ffd6c8b57ff24def866eb0213d 
>   3rdparty/libprocess/include/process/process.hpp c36df991b6a2c120ab0562e8ff907f9fbf8630d1

>   3rdparty/libprocess/src/CMakeLists.txt 0ce7dac5deea94623530820d173ce3ffe5b42ea4 
>   3rdparty/libprocess/src/memory_profiler.hpp PRE-CREATION 
>   3rdparty/libprocess/src/memory_profiler.cpp PRE-CREATION 
>   3rdparty/libprocess/src/process.cpp 9eb37465cd86f408d69f5f98fb76c4f4b93b9acd 
>   CHANGELOG c02d56d6612c432bb079a15fde84cb30d7455971 
> 
> 
> Diff: https://reviews.apache.org/r/63368/diff/11/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Benno Evers
> 
>


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