we have ambari cluster with 3 masters machines , two kafka’s and 3 workers machines
each worker have 5 disks as :
/dev/sdb 200G 477M 200G 3% /gr/sdb /dev/sdc 200G 183M 200G 1% /gr/sdc /dev/sdd 200G 283M 200G 2% /gr/sdd /dev/sde 200G 45M 200G 1% /gr/sde
we want to add additional 5 disks to each worker machine as ( /dev/sdf , /dev/sdh …, etc )
remark – our Workers nodes runs both a DataNode and NodeManager
for now I understand that the following parameters should be update according to the new disks
dfs.datanode.data.dir ( from HDFS ) yarn.nodemanager.local-dirs ( from YARN ) yarn.nodemanager.log-dirs ( from YARN )
regarding that do we missed something else? that need to be configured?
second
after we reconfigured the parameters in ambari GUI ,
how to verify that ambari cluster recognize the new disks on worker machines?
Advertisement
Answer
First, you need to partition the disks, then mount them in the /etc/fstab
to a directory.
Those settings look correct, though.
how to verify that ambari cluster recognize the new disks on worker machines
It will alert if the configured volumes in the list of those properties are not mounted.