hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From ESGLinux <esggru...@gmail.com>
Subject question about failover and fencing
Date Mon, 24 Dec 2012 09:39:44 GMT
Hi all,

I´m planning to use NN HA and I have a question about fencing:

when you run this command:

hdfs haadmin -failover namenode01 namenode02

the first namenode01 gets allways fenced? or simply gets inactive?

As far as I know the standby nanemode makes the checkpoint task for the
active namenode, so what happens when you fence a namenode, you only have a
namenode, who does the checkpoint task in this scenario?

Thanks in advance,

ESGLinux,

Mime
View raw message