flume-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Phil Scala <Phil.Sc...@globalrelay.net>
Subject RE: Spooling fileSuffix attribute ignored
Date Wed, 22 May 2013 22:55:54 GMT
Ah Mike, nice to see the ignore pattern ☺  that’s quite handy.

As for your question to me:
I am still using the 1.3.x version at the moment.  As for comments on the spooled directory
source, It still early in development but so far it has worked fairly well, though I have
not thrown too much at it, my biggest test was dropping 300 files to be parsed, and it chugged
through very nicely)

I am looking at maybe taking a crack at – https://issues.apache.org/jira/browse/FLUME-1899
sub directory support, we recently had some discussions and some of the design may change
and warrant implementing that.


One observation I have had from some failure tests, if I kill -9 my agent process mid file
processing, when my agent is restarted the spool source will basically start at the top of
that same file.  So there are chances for duplicates, and in my use this is good…but It
may not be an expectation to others.

I will definitely report back with findings as I move along, this is a fast moving project
so I will have some better “numbers”  very soon.

Thanks
  Phil


Phil Scala
Software Developer / Architect
Global Relay

phil.scala@globalrelay.net<mailto:phil.scala@globalrelay.net>

866.484.6630  |  info@globalrelay.net<https://ex1.globalrelay.net/owa/redir.aspx?C=01df0579a4cc42feb9f7e2801368f729&URL=mailto%3ainfo%40globalrelay.net>
 |  globalrelay.com<https://ex1.globalrelay.net/owa/redir.aspx?C=01df0579a4cc42feb9f7e2801368f729&URL=http%3a%2f%2fwww.globalrelay.com%2f>

From: Mike Percy [mailto:mpercy@apache.org]
Sent: Wednesday, May 22, 2013 3:59 AM
To: user@flume.apache.org
Subject: Re: Spooling fileSuffix attribute ignored

You should check whether your version of Flume supports the ignorePattern configuration param.
The latest version on trunk does.

Mike


On Wed, May 22, 2013 at 12:43 AM, ZORAIDA HIDALGO SANCHEZ <zoraida@tid.es<mailto:zoraida@tid.es>>
wrote:
Oh! I see, then was a miss understanding. Ok, so we will need to find a workaround. Thanks
a lot.

De: Mike Percy <mpercy@apache.org<mailto:mpercy@apache.org>>
Responder a: Flume User List <user@flume.apache.org<mailto:user@flume.apache.org>>
Fecha: miércoles, 22 de mayo de 2013 09:35
Para: Flume User List <user@flume.apache.org<mailto:user@flume.apache.org>>
Asunto: Re: Spooling fileSuffix attribute ignored

Hi Phil,
Nice approach. How is the spooling directory source working for you? Any thoughts on how it
could be improved?

Mike

On Tue, May 21, 2013 at 8:17 AM, Phil Scala <Phil.Scala@globalrelay.net<mailto:Phil.Scala@globalrelay.net>>
wrote:
Hi,

Based on my use and understanding that setting “fileSuffix” is simpy the extension to
the file to be added once the file was consumed and placed onto the channel.  I don’t think
it was intended to be used to indicate a completely uploaded file.    In the dev newsgroup
there was a discussion about having the spooler ”wait” for a little while before ingesting
the file.  An ignore pattern may also be a good idea.

Currently I am using an upload directory that I monitor and when lsof reports a file is not
in use/open I then move it to my spool directory.

HTH
Phil



Phil Scala
Software Developer / Architect
Global Relay

phil.scala@globalrelay.net<mailto:phil.scala@globalrelay.net>

866.484.6630<tel:866.484.6630>  |  info@globalrelay.net<https://ex1.globalrelay.net/owa/redir.aspx?C=01df0579a4cc42feb9f7e2801368f729&URL=mailto%3ainfo%40globalrelay.net>
 |  globalrelay.com<https://ex1.globalrelay.net/owa/redir.aspx?C=01df0579a4cc42feb9f7e2801368f729&URL=http%3a%2f%2fwww.globalrelay.com%2f>

From: ZORAIDA HIDALGO SANCHEZ [mailto:zoraida@tid.es<mailto:zoraida@tid.es>]
Sent: Tuesday, May 21, 2013 9:21 AM
To: Flume User List
Subject: Spooling fileSuffix attribute ignored

Dear all,
I am using the Spooling attribute "fileSuffix" to ensure that only when a file has been completely
upload into the Spooling dir,  is consumed by the source. However, files with no suffix are
also appended to the channel and then processed by the sink.
My configuration:

tier1.sources  = s1
tier1.channels = c1
tier1.sinks    = s1

# For each source, channel, and sink, set
# standard properties.
tier1.sources.s1.type     = spooldir
tier1.sources.s1.spoolDir = /home/user/flume/data
tier1.sources.s1.deletePolicy = immediate
tier1.sources.s1.batchSize = 1000
tier1.sources.s1.bufferMaxLines = 3000
tier1.sources.s1.fileHeader = true
tier1.sources.s1.fileSuffix=.COMPLETED

Is that ok?

Thanks.

________________________________

Este mensaje se dirige exclusivamente a su destinatario. Puede consultar nuestra política
de envío y recepción de correo electrónico en el enlace situado más abajo.
This message is intended exclusively for its addressee. We only send and receive email on
the basis of the terms set out at:
http://www.tid.es/ES/PAGINAS/disclaimer.aspx


________________________________

Este mensaje se dirige exclusivamente a su destinatario. Puede consultar nuestra política
de envío y recepción de correo electrónico en el enlace situado más abajo.
This message is intended exclusively for its addressee. We only send and receive email on
the basis of the terms set out at:
http://www.tid.es/ES/PAGINAS/disclaimer.aspx

Mime
View raw message