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

Switch to Threaded View
Flume, mail # user - hdfs.idleTimeout ,what's it used for ?


Copy link to this message
-
Re: hdfs.idleTimeout ,what's it used for ?
Juhani Connolly 2013-01-18, 03:39
That breaks the use case idleTimeout was originally made for: making
sure the file is closed promptly after data stops arriving. We use this
to make sure the files ready for our batches which run quite soon after.
The time that rollInterval will trigger is unpredictable as it will
reset every time any other type of roll is triggered(event count or size).

By making rollInterval behave properly all of this is a non-issue. My
recommendation to users woudl be not to use rollInterval if they're
bucketing by time(it's redundant behavior).

Documentation could definitely be improved. Once we sort out the
approach we want to take I can write it up to make the difference and
usage clearer.

On 01/18/2013 12:24 PM, Connor Woodson wrote:
> The way idleTimeout works right now is that it's another rollInterval;
> it will work best when rollInterval is not set and so it seems that
> it's use is best for when you don't want to use a rollInterval and
> just want to have your bucketwriters close when no events are coming
> through (caused by path change or something else; and you can still
> roll reliably with either count or size)
>
> As such, perhaps it is more clear if idleTimeout is renamed to
> idleRoll or such?
>
> And then change idleTimeout to only count seconds since it was closed;
> if a bucketwriter is closed for long enough it will automatically
> remove itself. This type of idle will then work well with
> rollInterval, while the other one doesn't (idleRoll + rollInterval
> creates two time-based rollers. There are certainly times for that,
> but not all of the time).
>
> - Connor
>
>
> On Thu, Jan 17, 2013 at 6:46 PM, Juhani Connolly
> <[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>> wrote:
>
>     It seemed neater at the time. It's only an issue because
>     rollInterval doesn't remove the entry in sfWriters. We could
>     change it so that close doesn't cancel it, and have it check
>     whether or not the writer is already closed, but that'd be kind of
>     ugly.
>
>     @Mohit:
>
>     When flume dies unexpectedly the .tmp file remains. When it
>     restarts there is some logic in HDFS sink to recover it(and
>     continue writing from there). I'm not actually sure of the
>     specifics. You may want to try and just kill -9 a running flume
>     process on a test machine and then start it up, look at the logs
>     and see what happens with the output.
>
>     If flume dies cleanly the file is properly closed.
>
>
>     On 01/18/2013 11:23 AM, Connor Woodson wrote:
>>     And @ my aside: I hadn't realized that the idleTimeout is
>>     canceled by the rollInterval occurring. That's annoying. So
>>     setting a lower idleTimeout, and drastically decreasing
>>     maxOpenFiles to at most 2 * possible open files, is probably
>>     necessary.
>>
>>
>>     On Thu, Jan 17, 2013 at 6:20 PM, Connor Woodson
>>     <[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>> wrote:
>>
>>         @Mohit:
>>
>>         For the HDFS Sink, the tmp files are placed based on the
>>         hadoop.tmp.dir property. The default location is
>>         /tmp/hadoop-${user.name <http://user.name>} To change this
>>         you can add -Dhadoop.tmp.dir=<path> to your Flume command
>>         line call, or you can specify the property in the
>>         core-site.xml of wherever your HADOOP_HOME environment
>>         variable points to.
>>
>>         - Connor
>>
>>
>>         On Thu, Jan 17, 2013 at 6:19 PM, Connor Woodson
>>         <[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>> wrote:
>>
>>             Whether idleTimeout is lower or higher than rollInterval
>>             is a preference; set it before, and assume you get one
>>             message right on the turn of the hour, then you will have
>>             some part of that hour without any bucket writers; but if
>>             you get another message at the end of the hour, you will
>>             end up with two files instead of one. Set it idleTimeout