Support Questions

Find answers, ask questions, and share your expertise
Check out our newest addition to the community, the Cloudera Data Analytics (CDA) group hub.

ambari parameters - Ambari’s NameNode restart timeout parameter

we intend to perform ambari upgrade to 2.6.1

so we read the hortonworks doc -

and about the: ( from the doc )

For Large Clusters

"In a large cluster, starting the NameNode can take a long time. To ensure the Ambari requests to start the NameNode do not timeout during the upgrade, you should configure Ambari’s NameNode restart timeout parameter. To do this, add ten percent to the normal amount of time in seconds it takes to start your NameNode when restarted. That value should be used to set the upgrade.parameter.nn-restart.timeout in /etc/ambari-server/conf/ on the Ambari Server host. Then, restart Ambari Server before beginning the upgrade."

we try to find the "ambari’s NameNode restart timeout parameter

from the file: /etc/ambari-server/conf/ ( on ambari server machine )

but this parameters isn't exists

first question - what is the exactly name of this parameters?

second question - what is the amount of the workers in the cluster to set this parameters in the - ?

grep timeout  /etc/ambari-server/conf/

Super Mentor

Super Mentor

I am not sure about this parameter , but what we need is the parameter from the doc - the parameter that called - Ambari’s NameNode restart timeout parameter , if you think the doc talk about the parameter - upgrade.parameter.nn-restart.timeout then just confirm that so we can continue


Super Mentor

@Michael Bronson

This property is not there by default inside the Can you please try adding it as per the doc:

That value should be used to set the upgrade.parameter.nn-restart.timeout in /etc/ambari-server/conf/ on the Ambari Server host.  (followed by ambari server restart)<br>

ok I will add it but what is the value according to the cluster size ? , I mean the doc talk about to set this parameters in case we need to upgrade very large cluster system , so in that case what is the defintion of large cluster? - 10 workers? 20 workers 50 workers , and what is the value regarding that ?


Super Mentor

@Michael Bronson

EDIT: In Large clusters the NameNode startup can cause delay. This is because it might take long time to collect the block information's from the DataNodes and update the BP information (assuming thart in large clusters the data is also too high) . So you will need to set the value for this parameter based on the observations of your NameNode startup time. The value of this parameter can vary based on the amount of data that you have on your cluster and the Size of your cluster. There is no standard formula to decide the exact value as the NameNode startyup can be slow due to many other reasons as well like less memory (slow network/IO ..etc)

Suppose normally if your Active NameNode startup takes 30 Minutes then set the value of this parameter to slightly greater than 30 Minutes (like 35 Minutes) this will help in avoiding the timeout.

Take a Tour of the Community
Don't have an account?
Your experience may be limited. Sign in to explore more.