zookeeper-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mike Schilli...@perlmeister.com>
Subject Total Outage of all ZooKeeper Instances
Date Wed, 12 Oct 2011 07:04:58 GMT
Hi ZooKeeper folks,

we just had an outage with our application using ZooKeeper the 
other day and are investigating how it happened.

Our setup is 3 servers, each of them running

    * the application
    * a local ZooKeeper instance

All 3 ZooKeeper instances form a quorum. Each application talks to the
local ZooKeeper instance only.

Looking through the logs of our applications on the 3 different servers, 
it turns out that *all three* ZooKeeper instances returned -7 to their 
applications during the outage. The applications tried to throw away 
their handle and reconnect, but none of them was able to recover.

Anyone got any ideas on how to prevent this kind of problem from
happening?

Thanks for any help.

-- 
-- Mike

Mike Schilli
m@perlmeister.com

Mime
View raw message