You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@zookeeper.apache.org by GitBox <gi...@apache.org> on 2019/11/13 19:57:31 UTC

[GitHub] [zookeeper] symat commented on issue #1048: ZOOKEEPER-3188: Improve resilience to network

symat commented on issue #1048: ZOOKEEPER-3188: Improve resilience to network
URL: https://github.com/apache/zookeeper/pull/1048#issuecomment-553575587
 
 
   I was trying to reproduce with docker the behaviour mentioned by @anmolnar above. So far I haven't succeed but found an other bug: 
   
   When I disabled the 'actively used' ethernet interface of the current leader, the follower noticed. During a new leader election it tried to reconnect in parallel to all the registered election addresses of the old leader. Waiting for the connection attempt to fail on the unavailable address caused a timeout in the connection to the available address. This was a flaky situation, usually causing 2-3 subsequent leader elections, but after a 10-15 seconds the quorum become stable.
   
   I solved this by filtering the available hosts before trying to establish connections to the Leader in the Learners.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services