You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@accumulo.apache.org by GitBox <gi...@apache.org> on 2019/05/16 16:52:40 UTC

[GitHub] [accumulo] ctubbsii commented on a change in pull request #1164: Fix #1084 partial - IT tests failing on Standalone

ctubbsii commented on a change in pull request #1164: Fix #1084 partial - IT tests failing on Standalone
URL: https://github.com/apache/accumulo/pull/1164#discussion_r284802220
 
 

 ##########
 File path: TESTING.md
 ##########
 @@ -85,7 +85,8 @@ An alternative to the MiniAccumuloCluster for testing, a standalone Accumulo clu
 most tests. This requires a manual step of building and deploying the Accumulo cluster by hand. The build can then be
 configured to use this cluster instead of always starting a MiniAccumuloCluster.  Not all of the integration tests are
 good candidates to run against a standalone Accumulo cluster, these tests will still launch a MiniAccumuloCluster for
-their use.
+their use.  They also take longer to run and it may be necessary to provide a timeout.factor property; `-Dtimeout.factor=3`
+is recommended.
 
 Review comment:
   It's not clear which tests "They" refer to in "They also", or what they are being compared to in "longer to run".
   
   In general, it's not true that running tests against a standalone cluster take longer to run. Usually, they're faster, because the tests avoid starting up a new MiniAccumuloCluster.
   
   The `timeout.factor` has nothing to do with the standalone testing method. Rather, it has to do the performance of the machine on which the tests are being run. I would not recommend any specific value, since it depends so much on the user's test machine specs.
   
   Instead of documenting this in the standalone testing section, it'd be better to mention this property in a section that applies equally to standalone testing and normal builds. Also, it could merely mention that some tests may take longer on certain hardware, and it may be useful to scale the normal timeouts accordingly with the `timeout.factor`. A value of 3 could be used as an example, but I don't think it should be recommended. (On my laptop, with 8GB RAM, an SSD, and a 5th gen Intel i7 mobile CPU, the most I've ever needed was a timeout factor of 2.)

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services