ambari-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Robison <>
Subject RE: Redundant Ambari services
Date Mon, 09 May 2016 17:41:55 GMT
Thanks, unfortunately, because of our clients, we cannot deploy to VMs but are deploying to
physical machines. David

Best Regards,

David R Robison
Senior Systems Engineer

From: Andrew Stadtler []
Sent: Monday, May 9, 2016 1:40 PM
Subject: Re: Redundant Ambari services


The simple solution if you have an existing Virtual Machine infrastructure is to put the Ambari-Server
on a VM with HA that can be restarted automatically in the event of a hardware failure. This
usually works best if you move the database to some type of highly available cluster mysql,
postgres or oracle rac too.

On May 9, 2016, at 12:34 PM, David Robison <<>>

I am working on setting up a Hadoop cluster where we need to ensure no single point of failure.
As part of this, the question is how best to deploy the Ambari services (e.g. configuration
and monitoring) to provide automatic failover should one of the monitoring nodes fails. One
thought was to use something like corosync and pacemaker to start the ambri-server on the
failover server if the primary should fail. The other idea was to have the ambary-server running
on both servers and use a virtual IP with failover to automatically switch traffic from one
server to the other should one fail. We are deploying onto Ubuntu 14.04. Has anyone done anything
like this? Any thoughts on how to proceed? Thanks, David

David R Robison
Senior Systems Engineer
O. +1 512 247 3700
M. +1 512 608 3173<><>
Prometheus Security Group Global, Inc.
3019 Alvin Devane Boulevard
Building 4, Suite 450
Austin, TX 78741

View raw message