Migrating Apache Flume Flows to Apache NiFi: Log to HDFS

Migrating Apache Flume Flows to Apache NiFi:  Log to HDFS





I am moving some Apache Flume flows over to Apache NiFi, this is the first one I am doing.    I am grabbing log files from a directory and pushing over to HDFS.   We had added some features beyond what was available in the original Flume flow.

We also can visually see it running and get access to real-time metrics and stats via Apache NiFi's REST API.



Step-By-Step

  1. Enter your directory holding your logs or files.   Check permissions and set file regular expression.
  2. Update FetchFile properties.
  3. UpdateAttribute to change your file name to what you would like.
  4. PutHDFS to store you file to HDFS.
  5. Add notifications, alerts and failure handling.



ListFile



Fetch the files from the directory and then delete when done.


Other Attributes for ListFile

View the State of Files as They Are Read


Replace the FileName for HDFS Safe Saves

You can change the HDFS  Owner


It will create new directories



Configure Your HDFS Directory to Store To (can be many directories/subdirectories programmatically)



Set The Size of an HDFS Block


Choose Your Compression


Chose How You Want to Write Data (Append, Replace, Ignore, Fail, Append)



NiFi File Filter / Reg Ex To Match Input Name Patterns:   (wifi).*log

You can match any subset of files and directories that we may want to grab.
Also remove invalid characters from HDFS filename output

NiFi Output:

Attribute Values

absolute.path
/var/log/
file.creationTime
2019-08-12T10:08:53-0400
file.group
wheel
file.lastAccessTime
2019-08-12T10:08:53-0400
file.lastModifiedTime
2019-08-12T10:08:53-0400
file.owner
root
file.permissions
rw-r--r--
file.size
271413
filename
wifi-08-12-2019__10:08:53.107.log
path
./
uuid
6693546e-1241-4625-bbd8-b9dd5da9281a

Files Migrated
hdfs dfs -ls /flume
Found 7 items
-rw-r--r--   1 tspann hdfs     120824 2019-08-13 19:49 /flume/wifi-08-12-2019__100647.406.log
-rw-r--r--   1 tspann hdfs     125888 2019-08-13 19:49 /flume/wifi-08-12-2019__100651.863.log
-rw-r--r--   1 tspann hdfs     252917 2019-08-13 19:49 /flume/wifi-08-12-2019__100839.901.log
-rw-r--r--   1 tspann hdfs     255873 2019-08-13 19:49 /flume/wifi-08-12-2019__100845.958.log
-rw-r--r--   1 tspann hdfs     271315 2019-08-13 19:49 /flume/wifi-08-12-2019__100848.189.log
-rw-r--r--   1 tspann hdfs     271413 2019-08-13 19:49 /flume/wifi-08-12-2019__100853.107.log

-rw-r--r--   1 tspann hdfs      17078 2019-08-13 19:49 /flume/wifi.log


Source:

https://github.com/tspannhw/flume-to-nifi

Email Post