Hi,
In
https://hadoop.apache.org/docs/current/hadoop-project-dist/hadoop-hdfs/HdfsRollingUpgrade.html#Upgrade_with_Downtime,
it says following:
In a non-HA cluster, there are a NameNode (NN), a SecondaryNameNode (SNN) and
many DataNodes (DNs). The procedure for upgrading a non-HA cluster is similar
to upgrading an HA cluster except that Step 2 “Upgrade Active and Standby NNs”
is changed to below:
* Upgrade NN and SNN
* Shutdown SNN
* Shutdown and upgrade NN.
* Start NN with the “-rollingUpgrade
started<https://hadoop.apache.org/docs/current/hadoop-project-dist/hadoop-hdfs/HdfsRollingUpgrade.html#namenode_-rollingUpgrade>”
option.
* Upgrade and restart SNN
We have a HA cluster running with Hadoop 2 and we want to upgrade it to Hadoop
3.
Here are my questions:
1. Should we also same steps are the above if we want to perform upgrade
with downtime?
2. If yes, when restarting standby NN as in step 4 above, should we start it
with the “-rollingUpgrade
started<https://hadoop.apache.org/docs/current/hadoop-project-dist/hadoop-hdfs/HdfsRollingUpgrade.html#namenode_-rollingUpgrade>”
option. ?
My question is for step 4 when restarting SNN, do we also need start it with
‘-rollingUpgrade started’ option?
Thanks,
Jason
