aurora-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jordan Ly <>
Subject Review Request 63521: Fixed issue where saving attributes are not being persisted to log
Date Thu, 02 Nov 2017 18:12:05 GMT

This is an automatically generated e-mail. To reply, visit:

Review request for Aurora, David McLaughlin, Stephan Erb, and Bill Farner.

Repository: aurora


A bug was introduced when the old `MemAttributeStore` was revived. Previously, the `saveHostAttributes`
method did not return anything. However, after migrating to the DB stores, the signature of
the interface was changed to return a `boolean` if the save modified the previous attributes.
The new changes accidentally inverted the order. The old `MemAttributeStoreTest` did not explicity
test for this scenario so it went unnoticed.

The interface involved:
 * Save a host attribute in the attribute store.
 * @param hostAttributes The attribute we are going to save.
 * @return {@code true} if the operation changed the attributes stored for the given
 *         {@link IHostAttributes#getHost() host}, or {@code false} if the save was a no-op.
boolean saveHostAttributes(IHostAttributes hostAttributes);


  src/main/java/org/apache/aurora/scheduler/storage/mem/ 483af194787e967a97c908a62889233336407aba

  src/test/java/org/apache/aurora/scheduler/storage/ 34db54be6eecbf0eaeab4fa2a19e6a66469cea88



Fixed tests so they take into account the return value of the save.
Added a test explicity testing this behavior.


Jordan Ly

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