question Re: Storage space consideration for deploying AMS in Archives of Support Questions (Read Only)

Storage space consideration for deploying AMS

VSomani — Wed, 02 Dec 2015 07:39:27 GMT

What are the guidelines around allocating storage space for AMS service?

A partner is looking for ballpark number for a small / medium and large cluster so they can ensure that they don't run out of storage space for AMS

Re: Storage space consideration for deploying AMS

nsabharwal — Mon, 19 Aug 2019 12:44:55 GMT

Please see the above. You want to make sure that it's pointing to correct file system with enough space. I would say 200GB to 300GB dedicated disk is good start for small cluster. Many times, installation ends up on root thats why its important to pay attention to the above setting.

@vsomani

Re: Storage space consideration for deploying AMS

abajwa — Wed, 02 Dec 2015 08:43:36 GMT

There are some guidelines available in our doc too: http://docs.hortonworks.com/HDPDocuments/Ambari-2.1.2.1/bk_ambari_reference_guide/content/_ams_general_guidelines.html

Re: Storage space consideration for deploying AMS

VSomani — Thu, 03 Dec 2015 02:41:54 GMT

Thanks Neeraj

Re: Storage space consideration for deploying AMS

VSomani — Thu, 03 Dec 2015 02:42:22 GMT

Thank you Ali. This document will definitely be useful for partner. Will share it with them

Re: Storage space consideration for deploying AMS

abajwa — Thu, 03 Dec 2015 03:39:51 GMT

Thanks for asking the question...I learnt something new 🙂

Re: Storage space consideration for deploying AMS

dorio — Fri, 11 Dec 2015 05:46:53 GMT

While storage space is absolutely critical as @Neeraj Sabharwal and @Ali Bajwa wrote in their post we just "discovered" that also CPU is a key point. When HWX released AMS we began to deploy Ambari and AMS on the same machine, but soon the understood that for a production environment it could be a good practice to use one VM for Ambari and another VM for AMS, so the really high impact on computation resources of AMS didn't impact Ambari (sometimes, during the aggregation phase we got 16 CPU at 90% for 10/15 minutes).