-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42882/
-----------------------------------------------------------

Review request for Aurora, Bill Farner and Zameer Manji.


Repository: aurora


Description
-------

The solution is to do sub-selects instead of relying on a large join. For a 
given update this approach reduced the overall number of rows returned by h2 to 
mybatis from 531441 to just 81.


Diffs
-----

  src/jmh/java/org/apache/aurora/benchmark/JobUpdates.java 
0c9dbaee81d27bacee1b6bf51b33b51509e28a63 
  src/jmh/java/org/apache/aurora/benchmark/UpdateStoreBenchmarks.java 
1d8986ba117f32288510c299ea5634f90a9311e7 
  
src/main/resources/org/apache/aurora/scheduler/storage/db/JobUpdateDetailsMapper.xml
 fba7d4f39dffb0f431e957993b9ae7c03eba2c60 

Diff: https://reviews.apache.org/r/42882/diff/


Testing
-------

Before this change:
```
Benchmark                                           (instanceOverrides)   Mode  
Cnt   Score    Error  Units
UpdateStoreBenchmarks.JobInstructionsBenchmark.run                    1  thrpt  
  5  39.907 ±  4.694  ops/s
UpdateStoreBenchmarks.JobInstructionsBenchmark.run                   10  thrpt  
  5  24.641 ±  4.187  ops/s
UpdateStoreBenchmarks.JobInstructionsBenchmark.run                  100  thrpt  
  5   0.654 ±  0.080  ops/s
UpdateStoreBenchmarks.JobInstructionsBenchmark.run                 1000  thrpt  
  5   0.007 ±  0.001  ops/s
```

After this change:
```
Benchmark                                           (instanceOverrides)   Mode  
Cnt   Score    Error  Units
UpdateStoreBenchmarks.JobInstructionsBenchmark.run                    1  thrpt  
  5   41.838 ±  3.060  ops/s
UpdateStoreBenchmarks.JobInstructionsBenchmark.run                   10  thrpt  
  5   41.812 ±  2.373  ops/s
UpdateStoreBenchmarks.JobInstructionsBenchmark.run                  100  thrpt  
  5   38.281 ±  2.034  ops/s
UpdateStoreBenchmarks.JobInstructionsBenchmark.run                 1000  thrpt  
  5   20.459 ±  2.886  ops/s
```


Thanks,

Maxim Khutornenko

Reply via email to