Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Plain View
Flume >> mail # user >> flume tail source problem and performance


+
周梦想 2013-01-29, 07:24
+
Alexander Alten-Lorenz 2013-01-29, 07:29
+
Jeong-shik Jang 2013-01-29, 07:41
+
周梦想 2013-02-04, 07:27
Copy link to this message
-
Re: flume tail source problem and performance
Hi Andy,

1. "startFromEnd=true" in your source configuration means data missing
can happen at restart in tail side because flume will ignore any data
event generated during restart and start at the end all the time.
2. With agentSink, data duplication can happen due to ack delay from
master or at agent restart.

I think it is why Flume-NG doesn't support tail any more but does let
user handle using script or program; tailing is a tricky job.

My suggestion is to use agentBEChain in agent tier, and DFO in collector
tier; you can still lose some data during failover at failure.
To minimize loss and duplication, implementing checkpoint function in
tail also can help.

Having monitoring system to detecting failure is very important as well,
so that you can notice failure and do recovering reaction quickly.

-JS

On 2/4/13 4:27 PM, 周梦想 wrote:
> Hi  JS,
> We can't accept agentBESink. Because this logs are important for data
> analysis, we can't make any errors of the data. losing data,
> duplication are all not acceptable.
> one agent's configure is :
> tail("H:/game.log", startFromEnd=true) agentSink("hadoop48", 35853)
>
>
> every time this windows agent restart, it will resend all the data to
> collector server.
> if some reason we restart the agent node, we can't get the mark of log
> where the agent have sent.
>
>
> 2013/1/29 Jeong-shik Jang <[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>
>
>     Hi Andy,
>
>     As you set startFromEnd option true, resend might be caused by DFO
>     mechanism (agentDFOSink); when you restart flume node in DFO mode,
>     all events in different stages(logged, writing, sending and so on)
>     rolls back to logged stage, which means resending and duplication.
>
>     And, for better performance, you may want to use agentBESink
>     instead of agentDFOSink.
>     I recommend to use agentBEChain for failover in case of failure in
>     collector tier if you have multiple collectors.
>
>     -JS
>
>
>     On 1/29/13 4:29 PM, Alexander Alten-Lorenz wrote:
>
>         Hi,
>
>         you could use tail -F, but this depends on the external
>         source. Flume hasn't control about. You can write your own
>         script and include this.
>
>         What's the content of:
>         /tmp/flume/agent/agent*.*/ - directories? Are sent and sending
>         clean?
>
>         - Alex
>
>         On Jan 29, 2013, at 8:24 AM, 周梦想 <[EMAIL PROTECTED]
>         <mailto:[EMAIL PROTECTED]>> wrote:
>
>             hello,
>             1. I want to tail a log source and write it to hdfs. below
>             is configure:
>             config [ag1, tail("/home/zhouhh/game.log",startFromEnd=true),
>             agentDFOSink("hadoop48",35853) ;]
>             config [ag2, tail("/home/zhouhh/game.log",startFromEnd=true),
>             agentDFOSink("hadoop48",35853) ;]
>             config [co1, collectorSource( 35853 ),  [collectorSink(
>             "hdfs://hadoop48:54310/user/flume/%y%m/%d","%{host}-",5000,raw),collectorSink(
>             "hdfs://hadoop48:54310/user/flume/%y%m","%{host}-",10000,raw)]]
>
>
>             I found if I restart the agent node, it will resend the
>             content of game.log
>             to collector. There are some solutions to send logs from
>             where I haven't
>             sent before? Or I have to make a mark myself or remove the
>             logs manually
>             when restart the agent node?
>
>             2. I tested performance of flume, and found it's a bit slow.
>             if I using configure as above, there are only 50MB/minute.
>             I changed the configure to below:
>             ag1:tail("/home/zhouhh/game.log",startFromEnd=true)|batch(1000)
>             gzip
>             agentDFOSink("hadoop48",35853);
>
>             config [co1, collectorSource( 35853 ), [collectorSink(
>             "hdfs://hadoop48:54310/user/flume/%y%m/%d","%{host}-",5000,raw),collectorSink(
>             "hdfs://hadoop48:54310/user/flume/%y%m","%{host}-",10000,raw)]]
Jeong-shik Jang / [EMAIL PROTECTED]
Gruter, Inc., R&D Team Leader
www.gruter.com
Enjoy Connecting
+
周梦想 2013-02-04, 08:07
+
Jeong-shik Jang 2013-02-04, 08:13
+
GuoWei 2013-02-04, 11:46
+
周梦想 2013-02-06, 02:47
+
周梦想 2013-02-04, 07:33
+
Alexander Alten-Lorenz 2013-02-04, 07:39
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB