On Wed, Apr 9, 2014 at 12:54 PM, Pritchard, Charles X. -ND
charles.x.pritchard@disney.com wrote:
On Apr 9, 2014, at 8:06 AM, Brock Noland br...@cloudera.com wrote:
Hi Charles,
Exploring the idea of using append instead of creating new files with
HDFS every few minutes.
...
it's
Hi Charles,
Exploring the idea of using append instead of creating new files with
HDFS every few minutes.
I wonder if this is doable by setting rollCount to 0 and then using
rollInterval (or alternatively rollSize)?
There's certainly a history of append with HDFS, mainly, earlier
versions
When the channel is full, the JMSSource returns BACKOFF to
PollableSourceRunner which implements backoff logic which can be seen here:
https://github.com/apache/flume/blob/trunk/flume-ng-core/src/main/java/org/apache/flume/source/PollableSourceRunner.java#L139
On Mon, Mar 17, 2014 at 7:38 PM,
I think it should work to run the 1.4 JMS source in 1.3.
Out of curiosity, which vendor isn't shipping 1.4?
Brock
On Thu, Mar 13, 2014 at 4:43 PM, Christopher Shannon
cshannon...@gmail.comwrote:
I know that the JMS source is new in version 1.4, but has anyone tried
running this particular
transactionCapacity is a limit on events consumed or produced. We have
removed it in trunk.
On Wed, Mar 5, 2014 at 12:47 PM, Jimmy jimmyj...@gmail.com wrote:
perhaps strange question, but looking at the documentation for flume 1.4
under FILE CHANNEL
transactionCapacity 1000The maximum size
at 10:25 PM, Brock Noland br...@cloudera.com wrote:
Yes. Note that CDH3 is deprecated as CDH4 has recently had it's 5th
maintenance release (CDH 4.5).
On Sun, Feb 9, 2014 at 9:12 PM, Upender Nimbekar
nimbekar.upen...@gmail.com wrote:
Does Cloudera CDH3U6 (I know its old) come up with Start
Hi,
In the cast of no failures with a single source, single channel and single
sink you will see ordering. However, I believe when there is a failure file
channel will change ordering on rollback.
If strict ordering is required it's advisable to assign sequence numbers
upstream and then re-order
Flume itself just provides a script to start flume in the foreground.
Apache BigTop provides init scripts for flume.
On Feb 8, 2014 12:22 PM, Upender Nimbekar nimbekar.upen...@gmail.com
wrote:
Does anyone know how to stop and start flume using one script in clustered
environment. It looks there
How large is your heap?
You will likely want two data directories per disk. Also with a channel
that large I strongly recommend using back up checkpoints.
Additionally https://issues.apache.org/jira/browse/FLUME-2155 will be very
useful to you as well.
On Jan 30, 2014 4:21 AM, Umesh Telang
I am guessing you want to write a Spooling Directory Source deserializer.
http://flume.apache.org/FlumeUserGuide.html#event-deserializers
On Thu, Jan 30, 2014 at 3:58 AM, Chhaya Vishwakarma
chhaya.vishwaka...@lntinfotech.com wrote:
Hi,
I want to write my own custom source to handle
on ubuntu and what confused
me is not about the virtual address space but NIO Direct Memory increasing
boundlessly. It will finally cause the machine to report swap is low.
On Wed, Jan 22, 2014 at 1:08 AM, Brock Noland br...@cloudera.com wrote:
Looks like it's arena allocation. Basically it's
Looks like it's arena allocation. Basically it's nothing to worry about
since virtual address space on 64bit machines is not in short supply. You
could limit it with a parameter like so:
I'd check the hbase sink and serializer config. It's too quiet when it's
misconfigured.
On Wed, Jan 15, 2014 at 6:35 AM, Chhaya Vishwakarma
chhaya.vishwaka...@lntinfotech.com wrote:
Hi,
This is my configuration file for pushing data using flume to hbase
host1.sources = src1
Could you share some more information? For example:
* What version you are using
* Configuration file
* Log
On Wed, Jan 15, 2014 at 8:24 AM, Josh Marcus jmar...@meetup.com wrote:
Hey folks,
I didn't get any response at all to this, so let me ask more generally:
does automatic reload upon
+1
On Wed, Jan 8, 2014 at 12:56 PM, Otis Gospodnetic
otis.gospodne...@gmail.com wrote:
Hi,
Do others think it would be valuable to Flume to:
1) more clearly list all available sources, sinks, interceptors
Compare:
http://flume.apache.org/FlumeUserGuide.html
to:
On Thu, Jan 2, 2014 at 10:25 AM, Brock Noland br...@cloudera.com wrote:
On Tue, Dec 31, 2013 at 8:34 PM, ed edor...@gmail.com wrote:
Hello,
We are using Flume v1.4 to load JSON formatted log data into HDFS as
Avro. Our flume setup looks like this:
NXLog == (FlumeHTTPSource - HDFSSink w
: 412-256-8556 | www.rdx.com
On Thu, Jan 2, 2014 at 12:27 PM, Brock Noland br...@cloudera.com
wrote:
Jimmy, great to hear that method is working for you!
Devin, regarding the morphlines question. Since ML can have arbitrary
java plugins it *can* do just about anything. I generally
Yes, it is possible to handle multi-line events and handling stack traces
is very common place.
However, using exec source is going to be limiting. The correct solution
is:
1) Use spooling directory source
2) Write a little deserializer to handle your format.
Another solution is:
1) replace
Hi,
The error:
*java.lang.OutOfMemoryError: unable to create new native thread*
doesn't have anything todo with virtual address space in a 64bit process.
It's highly likely that your nproc setting is too low. I would increase
this. For example Cloudera Manager sets this to 32K which is a much
Please share your sink configuration.
On Thu, Dec 26, 2013 at 4:05 AM, Chhaya Vishwakarma
chhaya.vishwaka...@lntinfotech.com wrote:
Hi,
I want to parse a log file to get the highlighted text.
Regular expression used to parse
(\.\w+)(?=Exception)
But table in hbase shows no
Hi Devin,
Please find my response below.
On Wed, Dec 18, 2013 at 12:24 PM, Devin Suiter RDX dsui...@rdx.com wrote:
So, if I understand your position on sizing the source properly, you are
saying that the fsync operation is the costly part - it locks the device
it is flushing to until the
, 2013 at 2:23 PM, Brock Noland br...@cloudera.com wrote:
Hi Devin,
Please find my response below.
On Wed, Dec 18, 2013 at 12:24 PM, Devin Suiter RDX dsui...@rdx.comwrote:
So, if I understand your position on sizing the source properly, you are
saying that the fsync operation is the costly
I assume you can then configure rsyslog to use RELP when sending to Flume?
If do I think it'd be great.
On Dec 17, 2013 12:07 AM, Otis Gospodnetic otis.gospodne...@gmail.com
wrote:
Hi,
I was looking at Rsyslog and RELP the other day and wondering whether a
RELP Source would make sense for
(4), see agent2.conf. The results are the
same.
Thanks for looking into this,
Jan
On 11/14/2013 05:08 PM, Brock Noland wrote:
On Thu, Nov 14, 2013 at 2:50 AM, Jan Van Besien ja...@ngdata.com
mailto:ja...@ngdata.com wrote:
On 11/13/2013 03:04 PM, Brock Noland wrote
Same questions: Can you take and share 8-10 thread dumps while the sink is
taking events slowly?
Can you share your machine configuration?
On Dec 17, 2013 7:57 AM, David Sinclair dsincl...@chariotsolutions.com
wrote:
Hi,
I am using a File Channel connected to an AMQP Source and an HDFS Sink.
again.
thanks for looking into this.
On Tue, Dec 17, 2013 at 8:54 PM, Brock Noland br...@cloudera.com wrote:
Can you take and share 8-10 thread dumps while the sink is taking events
slowly?
Can you share your machine and file channel configuration?
On Dec 17, 2013 6:28 AM, Shangan Chen
Hi,
I'd also add the biggest issue I see with the file channel is batch size at
the source. Long story short is that file channel was written to guarantee
no data loss. In order to do that when a transaction is committed we need
to perform a fsync on the disk the transaction was written to.
Additionally I'd note that worrying about virtual memory on 64 bit machines
is probably not worth your time. The newer versions of malloc() do arena
allocation and reserve virtual memory for each thread. This does not
however, actually consume memory.
On Sat, Dec 14, 2013 at 10:49 AM, Matt Wise
service
On Tue, Nov 26, 2013 at 1:17 PM, ch huang justlo...@gmail.com wrote:
yes,that's what i wanted
On Tue, Nov 26, 2013 at 11:50 AM, Brock Noland br...@cloudera.com wrote:
It sounds like you want to run both the http and ganglia monitoring at
the same time. Is that correct?
On Monday
Sounds like the nproc ulimit...
On Tue, Nov 26, 2013 at 8:50 PM, Jeff Lord jl...@cloudera.com wrote:
Can you provide the logfile and config?
On Tue, Nov 26, 2013 at 12:20 PM, Cochran, David david.coch...@bsee.gov
wrote:
I've got a pretty good sized box collecting logs for a number of
On Tue, Nov 26, 2013 at 7:41 AM, GuoWei wei@wbkit.com wrote:
Dear all,
I use flume and custom base sink to put data to HBase. In flume, I use file
channel.
the file channel data put in the following folder.
razor.channels.c_error.dataDirs = /var/lib/flume-ng/data/error
But I see the
Lower the maxFileSize.
On Mon, Nov 25, 2013 at 2:41 PM, Ritesh Adval riteshad...@gaikai.com wrote:
Hi,
We are running two flume 1.4 agents each with 2 file channel on a VM of
size 15GB.
Is VM recommded to run flume or do we need bare metal boxes?
Every week or so we are running into
maxFileSize. Do we know
how many max log files it will keep in flume 1.4 ?
Ritesh
On Mon, Nov 25, 2013 at 12:50 PM, Brock Noland br...@cloudera.com wrote:
Lower the maxFileSize.
On Mon, Nov 25, 2013 at 2:41 PM, Ritesh Adval riteshad...@gaikai.com
wrote:
Hi,
We are running two
the channel capacity,
currently it is set to 1M
Ritesh
On Mon, Nov 25, 2013 at 1:00 PM, Brock Noland br...@cloudera.com wrote:
It will keep any tx log that has a corresponding event in the channel
+ 2 per data directory.
On Mon, Nov 25, 2013 at 2:55 PM, Ritesh Adval riteshad...@gaikai.com
It sounds like you want to run both the http and ganglia monitoring at the
same time. Is that correct?
On Monday, November 25, 2013, ch huang wrote:
hi,maillist:
i want to write a nagios plugin for flume monitor and
alert,the script need get input from the command
curl
On Thu, Nov 14, 2013 at 2:50 AM, Jan Van Besien ja...@ngdata.com wrote:
On 11/13/2013 03:04 PM, Brock Noland wrote:
The file channel uses a WAL which sits on disk. Each time an event is
committed an fsync is called to ensure that data is durable. Without
this fsync there is no durability
wrote:
Hi Brock,
Yes, I think morphline interceptor should be something I am looking for.
I am studying it now.
Thank you,
George
On Tue, Oct 29, 2013 at 12:56 PM, Brock Noland br...@cloudera.comwrote:
In a very simple demo you could use the static interceptor:
http
Under certain circumstances mv will actually be a copy + delete. In that
case the file size will change during the copy phase. I'd recommend copying
it in with an extension which is ignored via excludePattern and then
renaming it.
On Wed, Oct 30, 2013 at 12:49 PM, Snehal Nagmote
Congratulations!
On Wednesday, September 25, 2013, Juhani Connolly wrote:
**
Congratulations Roshan, nice job!
On 09/25/2013 02:52 PM, Arvind Prabhakar wrote:
Congratulations Roshan!
Regards,
Arvind Prabhakar
On Tue, Sep 24, 2013 at 4:05 PM, Mike Percy
reach the minimumRequiredSpace), together they use more than the default 2G
of this parameter.
On Tue, Sep 10, 2013 at 8:23 AM, Brock Noland br...@cloudera.com wrote:
If you are concerned about disk space consumption you should lower the
max log size on the file channel. The exact parameter
If you are concerned about disk space consumption you should lower the
max log size on the file channel. The exact parameter is in the docs.
On Tue, Sep 10, 2013 at 12:17 AM, Anat Rozenzon a...@viber.com wrote:
After leaving flume to run in this state (sink is not sending the events),
the disk
In your case I would look at the spooling directory source.
On Sun, Aug 18, 2013 at 9:29 PM, Wang, Yongkun | Yongkun | BDD
yongkun.w...@mail.rakuten.com wrote:
Hi,
I am testing with apache-flume-1.4.0-bin.
I made a naive python script for exec source to do throttling by calling
sleep()
use-fast-replay would help but you'd need 4-5GB of heap per channel. With
heaps that large you use be using dual checkpointing to avoid this.
Here is the thread doing the replay:
lifecycleSupervisor-1-0 prio=10 tid=0x7f040472c800 nid=0x1332b
runnable [0x7f03f84ce000]
collector.channels.mc1.transactionCapacity=1
collector.channels.mc1.use-fast-replay=true
On Thu, Aug 8, 2013 at 3:19 PM, Brock Noland br...@cloudera.com wrote:
use-fast-replay would help but you'd need 4-5GB of heap per channel. With
heaps that large you use be using dual checkpointing
it is.
Should I just set:
useDualCheckpoints=true
backupCheckpointDir=/some/dir
On Thu, Aug 8, 2013 at 4:05 PM, Brock Noland br...@cloudera.com wrote:
Note the dual and backup checkpoint configs here:
http://flume.apache.org/FlumeUserGuide.html#file-channel
On Thu, Aug 8, 2013 at 7:37
Weird... It could have been the linux oom killer? You'd see something in
/var/log/message if that was the case.
On Sat, Jun 1, 2013 at 2:47 AM, Lenin Raj emaille...@gmail.com wrote:
Hello,
I have a flume service which pulls twitter data and sinks to HDFS. I
started it last night at 8 PM. It
I'd be highly surprised if not setting batchSize caused the NPE. Here is a
good tuning article:
https://blogs.apache.org/flume/entry/flume_performance_tuning_part_1
On Mon, May 13, 2013 at 10:58 AM, Jay Vyas jayunit...@gmail.com wrote:
Alas - i cannot reproduce this bug anymore., maybe as a
You only need the hadoop-core jar. Thanks for pointing out 1137. Work is
progressing on this issue in
FLUME-1285https://issues.apache.org/jira/browse/FLUME-1285
.
On Thu, May 2, 2013 at 8:15 AM, brian.h...@bt.com wrote:
Hi I'm running up against FLUME-1137 (
Weird. Do you feel comfortable using a java profiler? If not could you take 4-5
thread dumps while that channel is using that much CPU.
--
Brock Noland
Sent with Sparrow (http://www.sparrowmailapp.com/?sig)
On Monday, April 29, 2013 at 2:48 AM, Pranav Sharma wrote:
Hi,
I see a very
Also, the below doesn't make sense unless you are generating the config
yourself. The file channel doesn't do the kind of path manipulation that
the hdfs sink does...
collector.channels.fileChannel.checkpointDir=~/.flume/file-channel/checkpoint_%{data_type}
Hi,
There are two issues with your configuration:
1) batch size of 1 with file channel is anti-pattern. This will result in
extremely poor performance because the file channel will have to do an
fsync() (expensive disk operation required to ensure no data loss) for each
event. Your batch size
I think we just need to copy 4-5 classes over from hadoop. Someone uploaded
a patch I just haven't had time to work with it.
On Wed, Apr 10, 2013 at 3:41 PM, Hari Shreedharan hshreedha...@cloudera.com
wrote:
Yes. We will need to remove the format upgrade code and the old format
code from
, Brock Noland br...@cloudera.com wrote:
There is no harm in setting write-timeout to something like 30 seconds. In
fact it probably makes sense to increase the default to 30 seconds.
On Mon, Apr 8, 2013 at 1:38 PM, Madhu Gmail madhu.munag...@gmail.comwrote:
Hello,
** **
I am getting
How large is /local/flume/file-channel/flume-log-sink-dev/data/log-884? Would
you be willing to share the file with me (off list) so I could take a look
at the corruption?
Brock
On Fri, Mar 29, 2013 at 1:02 PM, Andrew Jones andrew+fl...@andrew-jones.com
wrote:
Hi,
I restarted my flume
Strange... Can you share 1) JVM 2) OS and 3) Flume config file?
On Thu, Mar 21, 2013 at 7:54 AM, JR mailj...@gmail.com wrote:
Hello,
Am a newbie, trying to set up acvro source, channel memory and hdfs
sink. Am encountering the following error: Not sure
what to do.
*Is there a way
Try turning on HeapDumpOnOutOfMemoryError so we can peek at the heap dump.
--
Brock Noland
Sent with Sparrow (http://www.sparrowmailapp.com/?sig)
On Friday, March 1, 2013 at 5:57 PM, Denis Lowe wrote:
process failed - java.lang.OutOfMemoryError
We observed the following error:
01 Mar
move a tremendous amount
of data between clusters and use compression and batching.
https://issues.cloudera.org/browse/FLUME-559
looks like I will need to spend time working with both versions.
Jim
From: Brock Noland br...@cloudera.com
Reply-To: user@flume.apache.org
Date: Sat, 23
Hi,
The spool dir source does not have built in functionality to read
compressed files. One could be built, but I think it would either require a
subclass of
https://github.com/apache/flume/blob/trunk/flume-ng-core/src/main/java/org/apache/flume/serialization/EventDeserializer.javaor
changes to
Can you share more about your tests? How many events you sent? How those times
were calculated, etc.
--
Brock Noland
Sent with Sparrow (http://www.sparrowmailapp.com/?sig)
On Wednesday, February 20, 2013 at 9:42 PM, Ryo.Hongo wrote:
Hi.
I did benchmark of the flume. 3〜8ms in one
Hi,
The spooling fir source expects immutable, uniquely named files
as described here:
http://flume.apache.org/FlumeUserGuide.html#spooling-directory-source
As such you should log to a separate directory and then on roll move the
file (uniquely named) into the spooling dir source.
Brock
On
Hi,
Try doing a mvn install -DskipTests before executing your commands.
On Mon, Feb 18, 2013 at 4:14 AM, 周梦想 abloz...@gmail.com wrote:
hello,
my env:
[zhouhh@Hadoop48 flume]$ cat /etc/redhat-release
CentOS release 5.9 (Final)
[zhouhh@Hadoop48 flume]$ uname -a
Linux Hadoop48
Hi,
Check to make sure you spelled your principal correctly and that your key
tab has the correct permissions.
On Thursday, February 14, 2013, Madhusudhan Reddy Munagala wrote:
Hello,
I am trying to configure the FlumeNG with HBase. We have Zoo keeper
service for the Hbase.
The Flume
NG has no zookeeper integration. What are you trying to do? There is one
open JIRA on this https://issues.apache.org/jira/browse/FLUME-1491
On Thu, Feb 7, 2013 at 8:36 PM, 吳瑞琳 rlwu...@gmail.com wrote:
Hi all,
I am trying to integrate Flume NG and zookeeper. However, I did not find
any
See this discussion:
http://mail-archives.apache.org/mod_mbox/flume-user/201301.mbox/%3C7475E65732997042ABDCE90B7B4EB286091644%40OZWEX0201N2.msad.ms.com%3E
On Mon, Feb 4, 2013 at 10:28 AM, Rahul Ravindran rahu...@yahoo.com wrote:
Re..sending.
--
*From:* Rahul
Hi,
That isn't a good error message. Could you share your entire log file? Post
it on a JIRA or pastebin.
Did you delete the checkpoint before starting the channel after changing
capacity? Capacity is fixed for the file channel as such the checkpoint
must be deleted to change the capacity.
takes a long time but succeeds. TestHBaseSink, however,
fails after a while when it times out.
How can I get this to work without running in 'sudo' mode, and why might
the TestHBaseSink be hanging for just me?
- Connor
On Sat, Jan 19, 2013 at 3:06 PM, Brock Noland br...@cloudera.com wrote
Also, Mike Percy has a working example here:
https://github.com/mpercy/flume-log4j-example
On Fri, Jan 25, 2013 at 11:10 AM, Hari Shreedharan
hshreedha...@cloudera.com wrote:
Have you set up the appender correctly? The log4j appender class is
I think there is/was a bug in HDFS which caused a NPE due to umask.
My guess is it's 0002 where as it needs to be 0022.
On Sat, Jan 19, 2013 at 2:56 PM, Connor Woodson cwoodson@gmail.com wrote:
Running mvn test on the latest Flume code, I get a test failure in
Hi,
Do I understand this correctly, you are going to use the flume log4j
appender to collect flume logs?
If so, I don't see how you'd avoid the feedback loop.
Brock
On Fri, Jan 18, 2013 at 11:13 AM, Connor Woodson cwoodson@gmail.com wrote:
I just ran into an unfortunate configuration
Hi,
Would you mind turning logging to debug and then posting your full log/config?
Brock
On Thu, Jan 17, 2013 at 8:24 PM, Henry Ma henry.ma.1...@gmail.com wrote:
Hi,
When using Spooling Directory Source in Flume NG 1.3.1, this exception
happens:
13/01/18 11:37:09 ERROR
, Brock Noland br...@cloudera.com wrote:
Hi,
Would you mind turning logging to debug and then posting your full
log/config?
Brock
On Thu, Jan 17, 2013 at 8:24 PM, Henry Ma henry.ma.1...@gmail.com wrote:
Hi,
When using Spooling Directory Source in Flume NG 1.3.1, this exception
happens
Hi,
I would use memory channel for now as opposed to file channel. For
file channel to keep up with that you'd need multiple disks. Also your
checkpoint period is super-low which will cause lots of checkpoints
and slow things down.
However, I think the biggest issue is probably batch size. With
.
I'm still not doing so great though. I'm getting about 300 Mb per minute in
my HDFS files. I should be getting about 300G. That's better than before
though. I've got 10% of the data this time, rather than 0.14% :)
On Jan 16, 2013, at 4:36 PM, Brock Noland br...@cloudera.com wrote
/debug$ zcat
collector102.ngpipes.sac.ngmoco.com.1358204406896.gz | wc -l
gzip: collector102.ngpipes.sac.ngmoco.com.1358204406896.gz: decompression
OK, trailing garbage ignored
100
This should be about 50,000 events for the 5 min window!!
Sagar
On Mon, Jan 14, 2013 at 3:16 PM, Brock Noland
Roy,
You deleted the checkpoint files AND the inflights and you got the same error?
Brock
On Mon, Jan 14, 2013 at 4:10 PM, Camp, Roy rc...@ebay.com wrote:
I somehow got one of my instances into a bad state where I continue to get
the following error. I have two data log files with about 2GB
This is https://issues.apache.org/jira/browse/FLUME-1779
On Mon, Jan 14, 2013 at 4:25 PM, Connor Woodson cwoodson@gmail.com wrote:
Oh alright, found it. What is happening is that the HDFS sink does not throw
an exception for this write error, but instead returns a Status.BACKOFF, and
as
[Log-BackgroundWorker-channel2] INFO
org.apache.flume.channel.file.LogFileV3 - Updating log-35.meta
currentPosition = 217919486, logWriteOrderID = 1358209947324
Sagar
On Mon, Jan 14, 2013 at 3:38 PM, Brock Noland br...@cloudera.com wrote:
Hmm, could you try and updated version of Hadoop
Hi,
OK. I would increase the capacity of the channel to say 200
with the original unmodified files.
I would also upgrade to the latest 1.3.1 since there are many file
channel fixes in 1.3.0 and 1.3.1.
On Mon, Jan 14, 2013 at 5:33 PM, Camp, Roy rc...@ebay.com wrote:
When I deleted both,
Nice post!
On Fri, Jan 11, 2013 at 12:13 PM, Mike Percy mpe...@apache.org wrote:
Hi folks,
I just posted to the Apache blog on how to do performance tuning with Flume.
I plan on following it up with a post about using the Flume monitoring
capabilities while tuning. Feedback is welcome.
payloadColumn or incrementColumn must be specified. This should give
you a better message and the open JIRA is
https://issues.apache.org/jira/browse/FLUME-1757
Brock
On Tue, Jan 8, 2013 at 10:19 PM, Xu (Simon) Chen xche...@gmail.com wrote:
Hi folks,
I am using flume to sink into hbase. I had
I would subclass HbaseEventSerializer for this purpose.
On Mon, Jan 7, 2013 at 10:33 AM, Xu (Simon) Chen xche...@gmail.com wrote:
Hi all,
Which hbase sink keeps the event header, and how to configure it to do so?
The SimpleHbaseEventSerializer would certainly discard the header:
I wouldn't modify the files while flume is also modifying them. It
might work but also might be a complete mess. If you need to modify
the events before being written interceptors are the correct solution.
After the file is done from a flume perspective, modify all you wish!
On Fri, Dec 21, 2012
Just as a side note: File Channel has received significantly more
testing than JDBC Channel.
On Wed, Dec 19, 2012 at 8:05 AM, Alexander Alten-Lorenz
wget.n...@gmail.com wrote:
Yes, why not luring the wiki :)
https://cwiki.apache.org/confluence/display/FLUME/Home
Stable:
whether the patch in FLUME-1794 fixes this.
Thanks,
Rudolf
-Original Message-
From: Brock Noland [mailto:br...@cloudera.com]
Sent: Tuesday, December 18, 2012 10:09 PM
To: user@flume.apache.org
Subject: Re: Flume 1.3.0 - NFS + File Channel Performance
Hi,
If you do have a chance
Hi,
Yes Flume 1.3. will be included in CDH 4.2.
Brock
On Wed, Dec 19, 2012 at 11:34 AM, Rahul Ravindran rahu...@yahoo.com wrote:
Hi,
Is Flume 1.3 part of CDH4? Is Flume 1.3 part of any debian repo for
installation? I have the link for http://flume.apache.org/download.html
which gives me
Hi,
I use the method mentioned on that page:
$ mkdir apache-flume
$ cd apache-flume
$ git clone https://git-wip-us.apache.org/repos/asf/flume.git
$ cd flume
$ mvn install -DskipTests
$ mvn eclipse:eclipse -DdownloadSources -DdownloadJavadocs
Add $HOME/.m2/repository to the my classpath in
Hi,
Hmm, yes in general performance is not going to be great over NFS, but
there haven't been any FC changes that stick out here.
Could you take 10 thread dumps of the agent running the file channel
and 10 thread dumps of the agent sending data to the agent with the
file channel? (You can
We'd need those thread dumps to help confirm but I bet that FLUME-1609
results in a NFS call on each operation on the channel.
If that is true, that would explain why it works well on local disk.
Brock
On Tue, Dec 18, 2012 at 10:17 AM, Brock Noland br...@cloudera.com wrote:
Hi,
Hmm, yes
.
--
Hari Shreedharan
On Tuesday, December 18, 2012 at 8:43 AM, Brock Noland wrote:
We'd need those thread dumps to help confirm but I bet that FLUME-1609
results in a NFS call on each operation on the channel.
If that is true, that would explain why it works well on local disk.
Brock
On Tue
Hi,
If you do have a chance, it would great to hear if the patch attached
to this JIRA (https://issues.apache.org/jira/browse/FLUME-1794) fixes
the performance problem.
Brock
On Tue, Dec 18, 2012 at 11:25 AM, Brock Noland br...@cloudera.com wrote:
Yeah I think we should do that check
Hi,
This is because Flume uses the NettyTransceiver and pyton avro only
supports HTTPTransciever.
This is not using avro, but you should be able to send JSON events to
the HTTPSource (http://flume.apache.org/FlumeUserGuide.html#http-source).
Brock
On Tue, Dec 18, 2012 at 3:13 PM, John Michaels
Hi,
FWIW, I was sending log data from Windows I would write a little
Windows Log Agent and send the data to the HTTP Source.
Brock
On Fri, Dec 14, 2012 at 8:47 AM, Kartashov, Andy andy.kartas...@mpac.ca wrote:
Flummers,
Loved working with Flume 1.2 – very easy and simple configuration, it
½[DATA_HERE]×ùÎ0ÆÜ9Ig::¬
;0
:[DATA_HERE]
I was hoping to avoid the and spaces (I'm assuming they're characters
that are encoded such that -cat won't show them).
Any thoughts?
Thanks again,
Chris
-Original Message-
From: Brock Noland [mailto:br...@cloudera.com]
Sent: Friday
Hi,
Why not try increasing the batch size on the source and sink to 10,000?
Brock
On Wed, Dec 12, 2012 at 4:08 AM, Jagadish Bihani
jagadish.bih...@pubmatic.com wrote:
I am using latest release of flume. (Flume 1.3.0) and hadoop 1.0.3.
On 12/12/2012 03:35 PM, Jagadish Bihani wrote:
Hi
Hi,
Just to echo Nitin, Flume 1.X is what the community is moving forward
with so you will likely find more support from the community on that
version. I'd highly recommend upgrading.
Cheers,
Brock
On Mon, Dec 10, 2012 at 3:13 AM, Nitin Pawar nitinpawar...@gmail.com wrote:
I just hope you are
Hi,
Hopefully someone will be able to answer the AVRO issue, in order to
help them, what version of Flume are you running?
Brock
On Thu, Dec 6, 2012 at 8:59 AM, DeCarlo, Thom tdeca...@mitre.org wrote:
Hi,
I'm just getting started with flume, so I apologize if this is an already
known
I should probably move my changes to my own local class, but this works for
now.
--
Thom DeCarlo
On Tue, 04 Dec 2012 17:15:06 GMT, Brock Noland br...@cloudera.com wrote:
Hi,
The host interceptor adds the host to the headers, not the body. You'd
need to create a new one to add
built as 1.4-SNAPSHOT.
--
Thom DeCarlo
-Original Message-
From: Brock Noland [mailto:br...@cloudera.com]
Sent: Thursday, December 06, 2012 10:06 AM
To: user@flume.apache.org
Subject: Re: AVRO_EVENT problem
Hi,
Hopefully someone will be able to answer the AVRO issue, in order
in the
JDBC Channel section.
--
Thom DeCarlo
-Original Message-
From: Brock Noland [mailto:br...@cloudera.com]
Sent: Thursday, December 06, 2012 1:57 PM
To: user@flume.apache.org
Subject: Re: AVRO_EVENT problem
It seems to me like the object you are trying to write doesn't match
OK, I don't really understand how Avro is working here, but I think
you should try FileChannel or maybe MemoryChannel for simplicity to
see if that works. IE, I think the problem is JDBCChannel. Can you let
me know how it turns out?
On Thu, Dec 6, 2012 at 1:45 PM, Brock Noland br...@cloudera.com
1 - 100 of 179 matches
Mail list logo