> On Oct. 2, 2012, 12:48 a.m., Hari Shreedharan wrote:
> > flume-ng-core/src/main/java/org/apache/flume/channel/MemoryChannel.java, 
> > lines 60-64
> > <https://reviews.apache.org/r/6982/diff/3/?file=155699#file155699line60>
> >
> >     Semaphores and synchronized/monitors are essentially different ways of 
> > achieving the same goal. 
> >     
> >     The only difference is that when you have permits > 1, semaphores are 
> > much better since each semaphore operation is atomic and you don't need to 
> > worry about handling thread context switches happening in between.
> >     
> >     Also in Java, traditionally java.util.concurrent.* classes have been 
> > implemented more efficiently than the language constructs themselves - 
> > though this seems to be fading over time, as synchornized/wait/notify have 
> > been getting better.
> 
> Hari Shreedharan wrote:
>     I intended to post this as reply to your query in the previous review. 
> Accidentally ended up posting it as a new review. Sorry about that.

So all you want is a semaphore permit = 1 blocking the adding of new events.  
That's easy to do, I just remember a review saying they didn't like the sync 
because the blocking would be too slow.  

So as I see it I can use the semaphores and have blocking or use the AtomicLong 
and error on the safe side of not blowing up.

Or in other words: Speed verse the rare chance of not getting the memory 
channel to it highest level of fullness.


- Ted


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/6982/#review12089
-----------------------------------------------------------


On Sept. 16, 2012, 3:54 a.m., Ted Malaska wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/6982/
> -----------------------------------------------------------
> 
> (Updated Sept. 16, 2012, 3:54 a.m.)
> 
> 
> Review request for Flume.
> 
> 
> Description
> -------
> 
> 1. The user will be able to define a byteCapacity and a 
> byteCapacityBufferPercentage.
> 2. Events byte size will be estimated from there body contents
> 3. On put bytes are added to current total
> 4. On commit any uncommitted takes are removed from the current byte total
> 5. On rollover any uncommitted puts are removed from the current byte total
> 
> 
> This addresses bug FLUME-1535.
>     https://issues.apache.org/jira/browse/FLUME-1535
> 
> 
> Diffs
> -----
> 
>   flume-ng-core/src/main/java/org/apache/flume/channel/MemoryChannel.java 
> c72e97c 
>   flume-ng-core/src/test/java/org/apache/flume/channel/TestMemoryChannel.java 
> e070864 
> 
> Diff: https://reviews.apache.org/r/6982/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Ted Malaska
> 
>

Reply via email to