GitHub user kl0u opened a pull request:
https://github.com/apache/flink/pull/2020
[FLINK-2314] Make Streaming File Sources Persistent
This PR solves FLINK-2314 and combines a number of sub-tasks. In addition,
it solves FLINK-3896 which was introduced as part of this task.
The way File Input sources are now processed is the following:
* One task monitors (parallelism 1) a user-specified path for new
files/data
* The above task assigns FileInputSplits to downstream (parallel)
readers to actually read the data
The monitoring entity scans the path, splits the files to be processed in
splits, and assigns them downstream. For now, two modes are supported. These
are the PROCESS_ONCE which just processes the current contents of the path and
exits, and the REPROCESS_WITH_APPENDED which periodically monitors the path and
reprocesses new files and (the entire contents of) files with new data.
In addition, these sources are checkpointed, i.e. in the case of a task
failure the job will resume from where it left off.
Finally, some changes were introduced in the way we are handling
FileInputFormats after discussions with @aljoscha .
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/kl0u/flink api_ft_files
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/2020.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #2020
----
commit 0d378f85ef7beec598701d78e0537f7479be99d9
Author: kl0u <kklou...@gmail.com>
Date: 2016-05-10T16:56:58Z
[FLINK-3896] Allow a StreamTask to be Externally Cancelled
It adds a method failExternally() to the StreamTask, so that custom
Operators
can make their containing task fail when needed.
commit 1a06e70d4cc72593663ed5065e9c68c5b9fadac1
Author: kl0u <kklou...@gmail.com>
Date: 2016-04-10T14:56:42Z
[FLINK-3717] Make FileInputFormat checkpointable
This adds a new interface called CheckpointableInputFormat
which describes input formats whose state is queryable,
i.e. getCurrentChannelState() returns where the reader is
in the underlying source, and they can resume reading from
a user-specified position.
This functionality is not yet leveraged by current readers.
commit 13db59ff214c6c1790b73e8b06984a7170924c5a
Author: kl0u <kklou...@gmail.com>
Date: 2016-04-18T14:37:54Z
[FLINK-3889] Refactor File Monitoring Source
This is meant to replace the different file
reading sources in Flink streaming. Now there is
one monitoring source with DOP 1 monitoring a
directory and assigning input split to downstream
readers.
In addition, it makes the new features added by
FLINK-3717 work together with the aforementioned entities
(the monitor and the readers) in order to have
fault tolerant file sources and exactly once guarantees.
This does not replace the old API calls. This
will be done in a future commit.
commit 0c8e852b96752a716c36452f7ced11c79cca5560
Author: kl0u <kklou...@gmail.com>
Date: 2016-05-18T14:44:45Z
[FLINK-2314] Make Streaming File Sources Persistent
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
