[
https://issues.apache.org/jira/browse/NIFI-1037?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
nicolas maillard updated NIFI-1037:
-----------------------------------
Description:
HDFS has an Inotify interface that enables to access the HDFS edit stream.
https://issues.apache.org/jira/browse/HDFS-6634
Creating a processor to listen in and get notifications either for select
directories or select actions would have many applications.
- Stream to a search engine the activity on HDFS
- Wait for specific actions or files to trigger workflows, like duplication to
other clusters
- Validate ingestion processes
etc..
probably more I don't think of.
I have a first working beta version that needs to evolve
it reuses the Hadoop-nar-bundle
Needs a HDFS 2.7 dependency currently done through editing the Hadop-lib bundle
let me know if this idea makes sense and would be of interest to the community
would love to contribute the idea
was:
HDFS has an Inotify interface that enables to access the HDFS edit stream.
https://issues.apache.org/jira/browse/HDFS-6634
Creating a processor to listen in and get notifications either for select
directories or select actions would have many applications.
- Stream to a search engine the activity on HDFS
- Wait for specific actions or files to trigger workflows, like duplication to
other clusters
- Validate ingestion processes
etc..
probably more I don't think of.
I have a first working beta version that needs to evolve
it reuses the Hadoop-nar-bundle
Needs a HDFS 2.7 dependency currently done through editing the Hadop-lib bundle
let me know if this idea makes sense and would be of interest to the community
> Hdfs Inotify processor
> ----------------------
>
> Key: NIFI-1037
> URL: https://issues.apache.org/jira/browse/NIFI-1037
> Project: Apache NiFi
> Issue Type: New Feature
> Reporter: nicolas maillard
> Priority: Minor
> Fix For: 0.4.0
>
>
> HDFS has an Inotify interface that enables to access the HDFS edit stream.
> https://issues.apache.org/jira/browse/HDFS-6634
> Creating a processor to listen in and get notifications either for select
> directories or select actions would have many applications.
> - Stream to a search engine the activity on HDFS
> - Wait for specific actions or files to trigger workflows, like duplication
> to other clusters
> - Validate ingestion processes
> etc..
> probably more I don't think of.
> I have a first working beta version that needs to evolve
> it reuses the Hadoop-nar-bundle
> Needs a HDFS 2.7 dependency currently done through editing the Hadop-lib
> bundle
> let me know if this idea makes sense and would be of interest to the community
> would love to contribute the idea
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
