Commit Graph

109 Commits

Author SHA1 Message Date
Xavier Léauté cd14e7fbca fix regex and assign random user ids 2013-09-05 13:43:25 -07:00
Xavier Léauté 5b6273ac3d fix null values in wikipedia decoder 2013-08-21 17:56:31 -07:00
Xavier Léauté 71b19af61f fix missing json property names 2013-08-21 17:38:30 -07:00
cheddar b4310d6cbf Merge pull request #219 from metamx/wikipedia-firehose
Wikipedia firehose
2013-08-21 14:34:07 -07:00
Xavier Léauté 7913b27066 reuse tmp file for geoip db 2013-08-21 14:12:39 -07:00
Xavier Léauté 0f4a132807 address code review 2013-08-21 13:57:50 -07:00
fjy ac32c6f27c fix tests according to code review 2013-08-21 13:23:38 -07:00
fjy 9c145f5ce3 basic unit tets for realtime functinality 2013-08-19 15:00:10 -07:00
Xavier Léauté 071e8cb2f3 updated javadocs 2013-08-15 11:07:55 -07:00
Xavier Léauté 3f33ca15b0 refactor namespaces and add docs / license 2013-08-15 10:58:03 -07:00
Gian Merlino dacee9ed20 Fix various InputRowParser serde issues 2013-08-14 16:51:02 -07:00
Xavier Léauté 6735ae4ecd add ip geo lookup 2013-08-14 14:29:44 -07:00
Xavier Léauté 1cc1d0be5f remove example code 2013-08-13 19:05:56 -07:00
cheddar c2305a82f9 Merge pull request #209 from zanox/kafka-protobuf
Kafka protobuf
2013-08-13 18:38:27 -07:00
Xavier Léauté 8ccac2f13d basic wikipedia firehose factory 2013-08-13 18:35:07 -07:00
Gian Merlino 1d1cb045f0 RealtimeManager: Handle FormattedExceptions at any stage of processing other than hasMore 2013-08-13 07:23:37 -07:00
Jan Rudert 3f32ef7bbc Merge branch 'master' of github.com:metamx/druid into kafka-protobuf 2013-08-12 17:42:40 +02:00
Gian Merlino ec53a30c6e Rejigger exception handling in standalone realtime
- Throw FormattedExceptions on null or unparseable timestamps
- Throw away events (and increment thrownAway) on non-FormattedExceptions in the fire chief
2013-08-06 17:32:08 -07:00
Jan Rudert 73d5ff4b94 Merge branch 'master' of github.com:metamx/druid into HEAD 2013-08-01 18:00:29 +02:00
Jan Rudert bba83ec532 introduced a ByteBufferInputRowParser to keep the old StringInputRowParser in places outside the KafkaFireHose 2013-07-29 11:59:08 +02:00
cheddar a6bc63732d Merge pull request #199 from metamx/indexing-service-mods
Improve robustness and usability of RealtimeIndexTask and some code cleanup in the Indexing Service
2013-07-25 17:09:15 -07:00
Gian Merlino 20acd6f450 RealtimePlumberSchool: Persist and merge immediately upon finishJob 2013-07-25 14:38:07 -07:00
fjy a909a31bcc better select sinks 2013-07-24 19:25:32 -07:00
Gian Merlino 5c96d500d8 More controllable realtime shutdown
- Realtime plumber will start persisting segments shortly after finishJob is called, regardless
  of rejection policy

- Add "none" rejection policy
2013-07-24 15:32:10 -07:00
Gian Merlino 0eebb0a149 Add RealtimeMetricsMonitor to RealtimeIndexTask 2013-07-24 14:39:59 -07:00
Gian Merlino 867dedfe1b Remove unused class 2013-07-24 10:27:21 -07:00
Gian Merlino 71704f7ee8 Replace graceful/hard shutdown combo with simple exit 2013-07-24 10:27:21 -07:00
fjy fc65648809 fix a bug introduced in postgres support that breaks inserts for mysql 2013-07-23 13:21:14 -07:00
cheddar 1c34e3f364 1) Fix class-level javadoc on RabbitMQFirehoseFactory to reflect changes in config 2013-07-19 16:47:00 -07:00
Jan Rudert 527c6c81c5 refactored StringInputRowParser to handle a ByteBuffer, better wiring and configuration of KafkaFirehose 2013-07-19 14:12:53 +02:00
Stefán Freyr Stefánsson ade00e441f Adding minor class-level JavaDoc. 2013-07-19 11:24:13 +00:00
Stefán Freyr Stefánsson af97b44a6f merging pull request from cheddar - jacksonifying the RabbitMQ connection and config. 2013-07-19 11:16:50 +00:00
cheddar 0edfbaffab Errr, needed to make the ConnectionFactory a JacksonifiedConnectionFactory 2013-07-18 13:17:11 -07:00
cheddar 807a52f963 Another take on the configuration 2013-07-18 13:16:42 -07:00
Stefán Freyr Stefánsson da658ac69a Changes according to Eric's comments on the pull request.
- moved the RabbitMQ ConnectionFactory configuration to a method of its own to clean up the FirehoseFactory.connect() method.
- using Throwables.propagate*() methods in exception handling.
- removed TODOs and instead using the "keyword" FUTURE as well as adding some context to the comments themselves.
- cleaned up the exception handling in hasMore() a little and made it more readable.
2013-07-18 19:30:14 +00:00
Stefán Freyr Stefánsson 22ec7c53fc Code cleanup and formatting, documentation and moving all configuration into the consumerProps. 2013-07-16 15:08:39 +00:00
Stefán Freyr Stefánsson 82d2623cc9 Minor sanity checks when checking for and processing a new message. 2013-07-15 15:25:59 +00:00
Stefán Freyr Stefánsson 13328a6b36 Implemented a better query.body for the test I'm using. Added a shutdown listener to the connection and channel to get a notification when they are closed. Question remains whether the connection should be re-established when that happens. 2013-07-15 09:50:58 +00:00
Stefán Freyr Stefánsson 4d998e4c27 Making more things configurable through the realtime.spec file. 2013-07-13 16:49:39 +00:00
Stefán Freyr Stefánsson 01b7507a5a First version of a RabbitMQ Firehose implementation. 2013-07-13 16:03:46 +00:00
Jan Rudert a8039107d8 no bytebuffer in the middle 2013-07-12 16:02:15 +02:00
Jan Rudert 89b0c84f3b initial implementation of an protocol buffers firehose 2013-07-12 13:34:57 +02:00
cheddar b42c5743c9 Merge pull request #177 from nebrera/master
Postgresql adaptation
2013-07-05 16:42:57 -07:00
cheddar 0712941825 1) Add check whether a Hydrant has already been persisted before persisting. Persisting happens synchronously on the same thread, but multiple persist requests can be queued up on that thread which means that subsequent ones would fail with an NPE. Fixes #178 2013-07-02 11:53:53 -07:00
Pablo Nebrera e9ce6b596c PostgresSQL integration with druid 2013-06-28 12:49:24 +02:00
fjy 15a4dae2b1 bunch of issue fixes 2013-06-26 10:45:54 -07:00
cheddar 9d80369c2b Merge branch 'master' of github.com:metamx/druid 2013-06-19 11:16:13 -07:00
cheddar 369412262e 1) Fix truncation math when figuring out what Segments to persist for realtime plumber. Previously, Sinks could get persist-n-merged without actually being ready to be persist-n-merged. Fixes #168 2013-06-19 11:15:58 -07:00
kOD-dKOD 270a4993ce Update RealtimePlumberSchool.java
Put a check to make sure that we're only looking for the folder those have number-only names, to avoid "merged" directory issue.
2013-06-11 12:05:17 +05:30
kOD-dKOD a0462374b0 Update RealtimePlumberSchool.java
There was an issue if you restart realtime node and merged dir is still there it will throw exceptions. To tackle this issue a check has been added. So now only persisted dir/files will be read and anything from "merged" will be avoided.
2013-06-07 14:02:25 +05:30