I implemented jt ha on cdh4.4.2 . Jobtracker keeps on failing over to each other, job keeps restarting, also namenode goes down at times and I can see logs for few datanodes mentioning all data nodes are bad. aborting.
I installed jt ha manually like this :-
After configuring jt ha i started jobtracker ha daemon using command after formatzk
Nohup Hadoop jobtrackerha &
Then i started mrzkfc using following commands
Nohup hadoop mrkfc &
Please advice me if I am doing anything wrong. Also is that right way to start the jt ha and failover controller ?
Sent from my iPad