Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Flume, mail # user - How to read multiples files getting continuously updated


Copy link to this message
-
RE: How to read multiples files getting continuously updated
Paul Chavez 2013-10-10, 16:04
That is exactly what I do for a similar scenario. In my case it's one big log file that gets written to all day on each server, so I developed a script that runs once a minute to grab the new lines off the file since last run, creates an incremental file with that data and then drops it in a directory for the spoolDir source to pick up.

Paul
From: Abhijeet Shipure [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, October 09, 2013 11:28 PM
To: [EMAIL PROTECTED]
Subject: Re: How to read multiples files getting continuously updated

Yes new files are created at fixed interval but write time is not fixed and files written as and when request comes.
I was thinking of creating utility to copy files to new directory and use Spool Dir source.

Regards
Abhijeet
On Thu, Oct 10, 2013 at 11:41 AM, Steve Morin <[EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>> wrote:
If the files are continually written to I don't think there is a good option.  Can new files be written to every time interval?

On Wed, Oct 9, 2013 at 11:09 PM, Abhijeet Shipure <[EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>> wrote:
Hi Steve,

Thanks for quick reply, as you pointed out Exec Source does not provide reliability, which is required in my case, and hence it is not suitable.

So which other inbuilt source could be used to read from many files ? Just one other requirement is file name s are also dynamically generated using time stamp after every 5 mins.
Regards
Abhijeet

On Thu, Oct 10, 2013 at 11:22 AM, Steve Morin <[EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>> wrote:
If your read the Flume manual it doesn't support a tail source

http://flume.apache.org/FlumeUserGuide.html#exec-source
Warning
The problem with ExecSource and other asynchronous sources is that the source can not guarantee that if there is a failure to put the event into the Channel the client knows about it. In such cases, the data will be lost. As a for instance, one of the most commonly requested features is the tail -F [file]-like use case where an application writes to a log file on disk and Flume tails the file, sending each line as an event. While this is possible, there's an obvious problem; what happens if the channel fills up and Flume can't send an event? Flume has no way of indicating to the application writing the log file that it needs to retain the log or that the event hasn't been sent, for some reason. If this doesn't make sense, you need only know this: Your application can never guarantee data has been received when using a unidirectional asynchronous interface such as ExecSource! As an extension of this warning - and to be completely clear - there is absolutely zero guarantee of event delivery when using this source. For stronger reliability guarantees, consider the Spooling Directory Source or direct integration with Flume via the SDK.
On Wed, Oct 9, 2013 at 10:33 PM, Abhijeet Shipure <[EMAIL PROTECTED]<mailto:[EMAIL PROTECTED]>> wrote:
Hi,

I am looking for Flume NG source that can be used for reading many files which are getting continuously updated.
I trued Spool Dir source but it does not work if file to be read gets modified.

Here is the scenario:
100 files are getting generated at one time and these files are continuously  updated for fixed interval say 5 mins, after 5 mins new 100 files get generated and being written again for 5 mins.

Which flume source is most suitable and how it should be used effectively without any data loss.

Any help is greatly appreciated.
Thanks
Abhijeet Shipure