accumulo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sean Busbey" <s...@manvsbeard.com>
Subject Re: Review Request 16908: Add a 3rd argument to functional test RunTests.java to allow a timeout factor for mapred.task.timeout and run.py -f flag
Date Wed, 15 Jan 2014 21:32:18 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/16908/#review31928
-----------------------------------------------------------



src/server/src/main/java/org/apache/accumulo/server/monitor/servlets/DefaultServlet.java
<https://reviews.apache.org/r/16908/#comment60679>

    Leave this change out, and put it in a follow on.



src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java
<https://reviews.apache.org/r/16908/#comment60680>

    1) you should do this in an overridden setup method, rather than on each run through map.
    
    2) Don't turn it into a file. Just use Configuration.getInt(key, default) (or the String
equivalent)



src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java
<https://reviews.apache.org/r/16908/#comment60681>

    If accumulo_timeout_factor is already a string, there's no reason call toString



src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java
<https://reviews.apache.org/r/16908/#comment60682>

    I would recommend not changing this to varargs



src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java
<https://reviews.apache.org/r/16908/#comment60684>

    You should define "accumulo.timeout_factor" as a constant in the class, to ensure consistency.
    
    Also, you should probably base it on the Class name to make sure it's unique.
    
    e.g.
    
      static final String TIMEOUT_FACTOR = RunTests.class.getName() + ".timeout_factor";



src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java
<https://reviews.apache.org/r/16908/#comment60683>

    likewise, I would not change this to varargs



src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java
<https://reviews.apache.org/r/16908/#comment60685>

    remove spurious String concatenation



test/system/auto/README
<https://reviews.apache.org/r/16908/#comment60686>

    suggest rewriting to:
    
    where 'timeout_factor' is an optional integer indicating how much we should scale up timeouts.
Will be used both to set mapred.task.timeout and to set hte "-f" flag on run.py.
    
    The 'timeout_factor' defaults to 1, which corresponds to a mapred.task.timeout of 480
seconds.


- Sean Busbey


On Jan. 15, 2014, 5:20 p.m., Hung Pham wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/16908/
> -----------------------------------------------------------
> 
> (Updated Jan. 15, 2014, 5:20 p.m.)
> 
> 
> Review request for accumulo.
> 
> 
> Bugs: ACCUMULO-2005
>     https://issues.apache.org/jira/browse/ACCUMULO-2005
> 
> 
> Repository: accumulo
> 
> 
> Description
> -------
> 
> ACCUMULO-2005: Add a 3rd argument to functional test RunTests.java to allow a timeout
factor for mapred.task.timeout and run.py -f flag.  Also changed 'ex' to 'e' on line 274 of
src/server/src/main/java/org/apache/accumulo/server/monitor/servlets/DefaultServlet.java
> 
> 
> Diffs
> -----
> 
>   src/server/src/main/java/org/apache/accumulo/server/monitor/servlets/DefaultServlet.java
eae0e81 
>   src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java 1a19bb5

>   test/system/auto/README 4773393 
> 
> Diff: https://reviews.apache.org/r/16908/diff/
> 
> 
> Testing
> -------
> 
> Testing was done on small cluster (4 nodes: master + 3 workers) to ensure MR job complete
successfully instead of failing out due to timeouts that lead to killed attempts.   Test results
can be provided upon request.
> 
> 
> Thanks,
> 
> Hung Pham
> 
>


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