kafka-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Shao, Saisai" <saisai.s...@intel.com>
Subject RE: Question about ConsoleConsumer
Date Thu, 17 Jul 2014 02:57:50 GMT
Hi Joe,

Thanks for your answer, seems I know what's its purpose.

Thanks a lot.
Jerry

-----Original Message-----
From: Joe Stein [mailto:joe.stein@stealth.ly] 
Sent: Thursday, July 17, 2014 10:15 AM
To: dev@kafka.apache.org
Cc: users@kafka.apache.org
Subject: Re: Question about ConsoleConsumer

auto.offset.reset is only when there is not an already established offset to consume from
or an offset that no longer exists (and is invalid) in the case of truncated logs or compaction).

You can force this behavior by deleting the znode

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 http://www.stealth.ly
 Twitter: @allthingshadoop <http://www.twitter.com/allthingshadoop>
********************************************/


On Wed, Jul 16, 2014 at 10:08 PM, Shao, Saisai <saisai.shao@intel.com>
wrote:

> Hi folks,
>
> A newbie question when investigating ConsoleConsumer's code,
>
> When we use ConsoleConsumer by setting "-from-beginning", we will set 
> property as "auto.offset.reset"="smallest", which means moving the 
> offset to the beginning of the data. But in the code, we still need to 
> delete Zookeeper's group metadata as:
>
>     if(options.has(resetBeginningOpt))
>       ZkUtils.maybeDeletePath(options.valueOf(zkConnectOpt), "/consumers/"
> + options.valueOf(groupIdOpt))
>
> From my understanding, setting "auto.offset.rest"="smallest" is enough 
> to seek to the beginning of the data, why we still need to delete 
> Zookeeper metadata? Is there any specific reason or some under-the-cover concerns?
>
> Sorry for the unthoughtful question and really appreciate your answer.
>
> Thanks
> Jerry
>
Mime
View raw message