----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/18626/#review36258 -----------------------------------------------------------
common/src/main/java/org/apache/falcon/catalog/AbstractCatalogService.java <https://reviews.apache.org/r/18626/#comment67209> Why do you need a separate method for this? Why not add this to getPartition method itself. Seems wasted IMO. common/src/main/java/org/apache/falcon/catalog/HiveCatalogService.java <https://reviews.apache.org/r/18626/#comment67210> Why do you need a separate method for this? Why not add this to getPartition method itself. Seems wasted IMO. Also, it naturally belongs to CatalogStorage, no? retention/src/main/java/org/apache/falcon/retention/FeedEvictor.java <https://reviews.apache.org/r/18626/#comment67212> nit: method can be map or fill rather than get retention/src/main/java/org/apache/falcon/retention/FeedEvictor.java <https://reviews.apache.org/r/18626/#comment67213> Can this method be on CatalogStorage? retention/src/main/java/org/apache/falcon/retention/FeedEvictor.java <https://reviews.apache.org/r/18626/#comment67215> nit: dropPartitionsForAnInstance retention/src/main/java/org/apache/falcon/retention/FeedEvictor.java <https://reviews.apache.org/r/18626/#comment67216> Why 2 booleans? dropped and deleted? rename dropped to deleted. - Seetharam Venkatesh On Feb. 28, 2014, 2:56 p.m., Satish Mittal wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/18626/ > ----------------------------------------------------------- > > (Updated Feb. 28, 2014, 2:56 p.m.) > > > Review request for Falcon and Srikanth Sundarrajan. > > > Repository: falcon-git > > > Description > ------- > > When an HCatalog based feed is scheduled in falcon, retention only looks at > the first partition key that satisfies either of date pattern: yyyy | MM | dd > | HH | mm. As a result, it calculates a partition filter that contains only > one of these patterns. However if HCatalog table is defined in such a way > that date spans across multiple partition keys (year/month/day/hour/minute), > then feed retention doesn't delete any partitions that are granular than > first level (year). > > > Diffs > ----- > > common/src/main/java/org/apache/falcon/catalog/AbstractCatalogService.java > fc9c3b1 > common/src/main/java/org/apache/falcon/catalog/HiveCatalogService.java > 3c3660e > common/src/main/java/org/apache/falcon/entity/common/FeedDataPath.java > 4031e14 > retention/src/main/java/org/apache/falcon/retention/FeedEvictor.java > 13c447c > > webapp/src/test/java/org/apache/falcon/lifecycle/TableStorageFeedEvictorIT.java > 770780e > > Diff: https://reviews.apache.org/r/18626/diff/ > > > Testing > ------- > > - Added new integration tests in TableStorageFeedEvictorIT.java to test > retention for an Hcatalog feed where date consists of multiple partitions > columns (year/month/day). > - Verified the retention behavior on a test cluster having an Hcatalog based > feed partitioned by year/month/day/hour/minute/country. > > > Thanks, > > Satish Mittal > >
