How do you add nodes to a Cloudbreak cluster?

[#1]

I have deployed a cluster in Azure using Cloudbreak and want to permanently add another two data nodes and, ideally, still have it managed through cloudbreak. How would I go about doing this?

[#2]

The easiest way is to use our shell. Please read details on wiki page.

[#3 · Accepted solution]

The best and easiest way to add new nodes to a cluster is if you using our cloudbreak shell which is available on github https://github.com/sequenceiq/cloudbreak-shell.

https://github.com/sequenceiq/cloudbreak-shell#upscaling-and-downscaling-a-stack

The version which working with cloudbreak.sequenceiq.com is 0.5.1: http://maven.sequenceiq.com/releases/com/sequenceiq/cloudbreak-shell/0.5.1/cloudbreak-shell-0.5.1.ja...

[#4]

On the cluster page you can enable autoscaling. From that point you have 2 options either set up time based or alert based scaling and you have to attach a scaling policy to it to know which host group you want to upscale.

The other option is the Cloudbreak cli/shell (helps you with TAB completion).

stack select --id <ID>

stack node ADD ..

(This feature is currently in Tech Preview).

[#5]

@khorvath@hortonworks.com

for those using https://cloudbreak.sequenceiq.com hosted version, I think only option is using a time based, right?

Would you have a simple example of "cron expression" usage to be executed only once and add few nodes to the cluster? I was showing cloudbreak to a partner and this was the first question: "how to add more nodes?"

[#6]

You can use metric based as well. We used to demo it to put RM WEB UI health to OK for 5 minutes, then add X nodes to slave_1 hostgroup. For cron: 0 54 10 ? * MON-FRI this would trigger at 10:54 Mon-Fri. Change the date according when you'd like to demo it.