Not sure if it's a ZK issue. Are you on ZK 3.3.3? Is the problem reproducible?
Thanks, Jun On Wed, Oct 12, 2011 at 10:52 AM, Mathias Herberts < mathias.herbe...@gmail.com> wrote: > > If that error occurs when the broker is down, this is normal. When the > > broker is brought up again, a rebalance will be triggered in the > consumers > > and a new fetcher to the restarted broker should be established. Could > you > > check if there is any rebalance after that error (search for > "rebalancing") > > and whether new fetchers are started afterwards? Also, what version of > Kafka > > are you using? > > Hi, > > the version of Kafka we are using is 0.6, what surprised me was the > fact that there was no rebalancing once the Kafka node came back up. > > Could it be a ZK issue with a missed watch? > > Mathias. >
