You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@kafka.apache.org by Meghana Narasimhan <mn...@bandwidth.com> on 2016/06/01 19:01:13 UTC

Restoring Kafka data to one broker

Hi,
I have a 3 node cluster with kafka version 0.9.0.1 with many topics having
replication factor 3 and isr 2.
Node 0 is running multiple mirrormakers.

Node 1 in this cluster has an issue and lost all data from disk (All kafka
data logs got deleted from disk).
Kafka broker is down but zookeeper is still running and its data folder is
available on this node.

Node 0 and 2 are still alive and processing data.

-Any inputs or suggestions on the best way to restore Node B and replicate
say 20TB of data from Node 0 and 2 ?
-Will just restarting the Kafka server replicate data to this node from
Node 0 and 2 ? (This seemed to work for a small volume of data, but
wondering if this is the right approach to take for larger volume of data)
-What are the best practices in terms of configuration parameters for
broker and mirrormaker that need to be verified in this scenario ?

Thanks,
Meghana

Re: Restoring Kafka data to one broker

Posted by Meghana Narasimhan <mn...@bandwidth.com>.
Hi All,
Any suggestions or inputs on this ? Any help would be greatly appreciated.

Thanks,
Meghana

On Wed, Jun 1, 2016 at 3:01 PM, Meghana Narasimhan <
mnarasimhan@bandwidth.com> wrote:

> Hi,
> I have a 3 node cluster with kafka version 0.9.0.1 with many topics having
> replication factor 3 and isr 2.
> Node 0 is running multiple mirrormakers.
>
> Node 1 in this cluster has an issue and lost all data from disk (All kafka
> data logs got deleted from disk).
> Kafka broker is down but zookeeper is still running and its data folder is
> available on this node.
>
> Node 0 and 2 are still alive and processing data.
>
> -Any inputs or suggestions on the best way to restore Node B and replicate
> say 20TB of data from Node 0 and 2 ?
> -Will just restarting the Kafka server replicate data to this node from
> Node 0 and 2 ? (This seemed to work for a small volume of data, but
> wondering if this is the right approach to take for larger volume of data)
> -What are the best practices in terms of configuration parameters for
> broker and mirrormaker that need to be verified in this scenario ?
>
> Thanks,
> Meghana
>
>