Re: Review Request 25133: Fixing race in task secondary index.

2014-08-28 Thread Kevin Sweeney

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25133/#review51823
---

Ship it!


Ship It!

- Kevin Sweeney


On Aug. 27, 2014, 6:19 p.m., Maxim Khutornenko wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://reviews.apache.org/r/25133/
 ---
 
 (Updated Aug. 27, 2014, 6:19 p.m.)
 
 
 Review request for Aurora and Kevin Sweeney.
 
 
 Bugs: AURORA-667
 https://issues.apache.org/jira/browse/AURORA-667
 
 
 Repository: aurora
 
 
 Description
 ---
 
 Synchronizing around index value retrieval. 
 
 Unit test is mostly verifying general functionality under multhithreaded 
 access. I was not able to repro this locally as doing so reliably would 
 require refactoring that would not make sense otherwise. 
 
 
 Diffs
 -
 
   src/main/java/org/apache/aurora/scheduler/storage/mem/MemTaskStore.java 
 f2656de4d723c35a2d3fbe75d9aeb99a353423a2 
   src/test/java/org/apache/aurora/scheduler/storage/mem/MemTaskStoreTest.java 
 6882c25399fb8c2e499f440ba205e5b0ba71d27a 
 
 Diff: https://reviews.apache.org/r/25133/diff/
 
 
 Testing
 ---
 
 gradle -Pq build
 
 
 Thanks,
 
 Maxim Khutornenko
 




Review Request 25133: Fixing race in task secondary index.

2014-08-27 Thread Maxim Khutornenko

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25133/
---

Review request for Aurora and Kevin Sweeney.


Bugs: AURORA-667
https://issues.apache.org/jira/browse/AURORA-667


Repository: aurora


Description
---

Synchronizing around index value retrieval. 

Unit test is mostly verifying general functionality under multhithreaded 
access. I was not able to repro this locally as doing so reliably would require 
refactoring that would not make sense otherwise. 


Diffs
-

  src/main/java/org/apache/aurora/scheduler/storage/mem/MemTaskStore.java 
f2656de4d723c35a2d3fbe75d9aeb99a353423a2 
  src/test/java/org/apache/aurora/scheduler/storage/mem/MemTaskStoreTest.java 
6882c25399fb8c2e499f440ba205e5b0ba71d27a 

Diff: https://reviews.apache.org/r/25133/diff/


Testing
---

gradle -Pq build


Thanks,

Maxim Khutornenko



Re: Review Request 25133: Fixing race in task secondary index.

2014-08-27 Thread Kevin Sweeney

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25133/#review51736
---



src/test/java/org/apache/aurora/scheduler/storage/mem/MemTaskStoreTest.java
https://reviews.apache.org/r/25133/#comment90297

This looks like it adds a time dependency to our unit tests, which might 
cause them to be flaky. Consider using a CountDownLatch instead.


- Kevin Sweeney


On Aug. 27, 2014, 6:19 p.m., Maxim Khutornenko wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://reviews.apache.org/r/25133/
 ---
 
 (Updated Aug. 27, 2014, 6:19 p.m.)
 
 
 Review request for Aurora and Kevin Sweeney.
 
 
 Bugs: AURORA-667
 https://issues.apache.org/jira/browse/AURORA-667
 
 
 Repository: aurora
 
 
 Description
 ---
 
 Synchronizing around index value retrieval. 
 
 Unit test is mostly verifying general functionality under multhithreaded 
 access. I was not able to repro this locally as doing so reliably would 
 require refactoring that would not make sense otherwise. 
 
 
 Diffs
 -
 
   src/main/java/org/apache/aurora/scheduler/storage/mem/MemTaskStore.java 
 f2656de4d723c35a2d3fbe75d9aeb99a353423a2 
   src/test/java/org/apache/aurora/scheduler/storage/mem/MemTaskStoreTest.java 
 6882c25399fb8c2e499f440ba205e5b0ba71d27a 
 
 Diff: https://reviews.apache.org/r/25133/diff/
 
 
 Testing
 ---
 
 gradle -Pq build
 
 
 Thanks,
 
 Maxim Khutornenko
 




Re: Review Request 25133: Fixing race in task secondary index.

2014-08-27 Thread Maxim Khutornenko


 On Aug. 28, 2014, 1:26 a.m., Kevin Sweeney wrote:
  src/test/java/org/apache/aurora/scheduler/storage/mem/MemTaskStoreTest.java,
   line 435
  https://reviews.apache.org/r/25133/diff/1/?file=670820#file670820line435
 
  This looks like it adds a time dependency to our unit tests, which 
  might cause them to be flaky. Consider using a CountDownLatch instead.

The CountdownLatch is already used to wait for completion (see above). This is 
simply to shut down executor cleanly and is the approach used elsewhere.


- Maxim


---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25133/#review51736
---


On Aug. 28, 2014, 1:19 a.m., Maxim Khutornenko wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://reviews.apache.org/r/25133/
 ---
 
 (Updated Aug. 28, 2014, 1:19 a.m.)
 
 
 Review request for Aurora and Kevin Sweeney.
 
 
 Bugs: AURORA-667
 https://issues.apache.org/jira/browse/AURORA-667
 
 
 Repository: aurora
 
 
 Description
 ---
 
 Synchronizing around index value retrieval. 
 
 Unit test is mostly verifying general functionality under multhithreaded 
 access. I was not able to repro this locally as doing so reliably would 
 require refactoring that would not make sense otherwise. 
 
 
 Diffs
 -
 
   src/main/java/org/apache/aurora/scheduler/storage/mem/MemTaskStore.java 
 f2656de4d723c35a2d3fbe75d9aeb99a353423a2 
   src/test/java/org/apache/aurora/scheduler/storage/mem/MemTaskStoreTest.java 
 6882c25399fb8c2e499f440ba205e5b0ba71d27a 
 
 Diff: https://reviews.apache.org/r/25133/diff/
 
 
 Testing
 ---
 
 gradle -Pq build
 
 
 Thanks,
 
 Maxim Khutornenko