You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@samza.apache.org by Boris Shkolnik <bo...@apache.org> on 2016/04/07 23:14:13 UTC

Re: Review Request 45504: SAMZA-924: Add disk space monitoring

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45504/#review127660
-----------------------------------------------------------


Ship it!




Please see the comment. Up to you to decide to act on it or not.


samza-core/src/main/scala/org/apache/samza/container/SamzaContainer.scala (line 398)
<https://reviews.apache.org/r/45504/#comment191047>

    Totally agree with your comment - it feels strange to hard code this stuff (although that what we do in TaskStorageManager). 
    May be it would be cleaner to add a getStoreLocation() method to the TaskStorageManager and call it for each taskInstance?


- Boris Shkolnik


On March 30, 2016, 7:43 p.m., Chris Pettitt wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45504/
> -----------------------------------------------------------
> 
> (Updated March 30, 2016, 7:43 p.m.)
> 
> 
> Review request for samza.
> 
> 
> Repository: samza
> 
> 
> Description
> -------
> 
> This change introduces the measurement of disk usage for selected
> directories (currently those used by stores only). The feature is off by
> default, but can be enabled by setting "container.disk.poll.interval.ms"
> to a non-zero value.
> 
> 
> Diffs
> -----
> 
>   samza-core/src/main/java/org/apache/samza/container/disk/DiskSpaceMonitor.java PRE-CREATION 
>   samza-core/src/main/java/org/apache/samza/container/disk/PollingScanDiskSpaceMonitor.java PRE-CREATION 
>   samza-core/src/main/scala/org/apache/samza/container/SamzaContainer.scala bcbc90a0a460f8733e6d3a50dbc33f3720cad7d0 
>   samza-core/src/main/scala/org/apache/samza/container/SamzaContainerMetrics.scala 6fae6509d177cc3a54dac9ad1d3e5cc479f4a4f5 
>   samza-core/src/test/java/org/apache/samza/container/disk/TestPollingScanDiskSpaceMonitor.java PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/45504/diff/
> 
> 
> Testing
> -------
> 
> - Added tests for our disk space monitoring implementation.
> - Verified metrics are correctly reported and updated when enabling the feature
> - Verified metrics are not attached to the container when the feature is disabled
> 
> Perf testing for this feature is still pending and is a requirement for
> this to be committed.
> 
> 
> Thanks,
> 
> Chris Pettitt
> 
>


Re: Review Request 45504: SAMZA-924: Add disk space monitoring

Posted by Chris Pettitt <cp...@linkedin.com>.

> On April 7, 2016, 9:14 p.m., Boris Shkolnik wrote:
> > samza-core/src/main/scala/org/apache/samza/container/SamzaContainer.scala, line 398
> > <https://reviews.apache.org/r/45504/diff/1/?file=1319902#file1319902line398>
> >
> >     Totally agree with your comment - it feels strange to hard code this stuff (although that what we do in TaskStorageManager). 
> >     May be it would be cleaner to add a getStoreLocation() method to the TaskStorageManager and call it for each taskInstance?

Thanks Boris. Actually that was not my comment. I moved this chunk out of a loop (a little further down in the diff) since we only need to evaluate it once.


- Chris


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45504/#review127660
-----------------------------------------------------------


On March 30, 2016, 7:43 p.m., Chris Pettitt wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45504/
> -----------------------------------------------------------
> 
> (Updated March 30, 2016, 7:43 p.m.)
> 
> 
> Review request for samza.
> 
> 
> Repository: samza
> 
> 
> Description
> -------
> 
> This change introduces the measurement of disk usage for selected
> directories (currently those used by stores only). The feature is off by
> default, but can be enabled by setting "container.disk.poll.interval.ms"
> to a non-zero value.
> 
> 
> Diffs
> -----
> 
>   samza-core/src/main/java/org/apache/samza/container/disk/DiskSpaceMonitor.java PRE-CREATION 
>   samza-core/src/main/java/org/apache/samza/container/disk/PollingScanDiskSpaceMonitor.java PRE-CREATION 
>   samza-core/src/main/scala/org/apache/samza/container/SamzaContainer.scala bcbc90a0a460f8733e6d3a50dbc33f3720cad7d0 
>   samza-core/src/main/scala/org/apache/samza/container/SamzaContainerMetrics.scala 6fae6509d177cc3a54dac9ad1d3e5cc479f4a4f5 
>   samza-core/src/test/java/org/apache/samza/container/disk/TestPollingScanDiskSpaceMonitor.java PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/45504/diff/
> 
> 
> Testing
> -------
> 
> - Added tests for our disk space monitoring implementation.
> - Verified metrics are correctly reported and updated when enabling the feature
> - Verified metrics are not attached to the container when the feature is disabled
> 
> Perf testing for this feature is still pending and is a requirement for
> this to be committed.
> 
> 
> Thanks,
> 
> Chris Pettitt
> 
>