You are viewing a plain text version of this content. The canonical link for it is here.

Posted to users@kafka.apache.org by Petri Lehtinen <pe...@digip.org> on 2015/12/30 10:51:25 UTC

Re: High CPU ~ Shrinking ISR for partition ~ skip updating ISR endless loop

Manish Sharma wrote:
> One of my brokers (id 0) is continuously emitting such log entries and
> eating up CPU cycles..
> 
> *[2015-11-07 12:49:50,677] INFO Partition [Wmt_Thursday_158,12] on broker
> 0: Shrinking ISR for partition [Wmt_Thursday_158,12] from 0,3 to 0
> (kafka.cluster.Partition)*
> 
> *[2015-11-07 12:49:50,680] INFO Partition [Wmt_Thursday_158,12] on broker
> 0: Cached zkVersion [2] not equal to that in zookeeper, skip updating ISR
> (kafka.cluster.Partition)*
> 
> *[2015-11-07 12:49:50,680] INFO Partition [Wmt_Saturday_139,8] on broker 0:
> Shrinking ISR for partition [Wmt_Saturday_139,8] from 0,8 to 0
> (kafka.cluster.Partition)*
> 
> *[2015-11-07 12:49:50,683] INFO Partition [Wmt_Saturday_139,8] on broker 0:
> Cached zkVersion [2] not equal to that in zookeeper, skip updating ISR
> (kafka.cluster.Partition)*

The same thing happened to me, too. There seems to be a related issue
in the issue tracker:

    https://issues.apache.org/jira/browse/KAFKA-3042