----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/13463/#review25018 -----------------------------------------------------------
In SpoolDirectorySource.shitdown(), it might be useful to add INFO or DEBUG level log indicating that it is waiting ... before entering a potentially 10 second long wait for the executor to wind up. Looks good overall. - Roshan Naik On Aug. 10, 2013, 2:33 a.m., Mike Percy wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/13463/ > ----------------------------------------------------------- > > (Updated Aug. 10, 2013, 2:33 a.m.) > > > Review request for Flume. > > > Bugs: FLUME-2157 > https://issues.apache.org/jira/browse/FLUME-2157 > > > Repository: flume-git > > > Description > ------- > > Spool directory source does not shut down correctly when Flume is > reconfigured. This can lead to all sorts of problems stemming from the fact > that we have threads running which should not be. In the logs, it looks like > multiple spool directory sources are running on the same directory, even > though the config may say otherwise. > > > Diffs > ----- > > > flume-ng-core/src/main/java/org/apache/flume/client/avro/ReliableSpoolingFileEventReader.java > f82fe1f > > flume-ng-core/src/main/java/org/apache/flume/source/SpoolDirectorySource.java > 7145580 > > flume-ng-core/src/test/java/org/apache/flume/source/TestSpoolDirectorySource.java > 652d2a2 > > flume-ng-tests/src/test/java/org/apache/flume/test/agent/TestSpooldirSource.java > PRE-CREATION > flume-ng-tests/src/test/java/org/apache/flume/test/util/StagedInstall.java > bc58340 > > Diff: https://reviews.apache.org/r/13463/diff/ > > > Testing > ------- > > Added a unit test for this specific case and also an integration test that > simply runs an agent with many concurrent spool directory sources. > > > Thanks, > > Mike Percy > >
