Can I assume you are using deduplication in combination with replication?
What is your expected daily ingest and how much managed data will you have
in the TSM server?

64GB CAN be enough for small (<1-2TB ingest) with replication to function
(not optimal) but for larger environments you quickly need to go higher.
Put your DB and active log on SSD's, don't mess around with 15K disks, they
work but don't scale well and you will add so many that SSD's are the way
to go.
You can use near line disks for filepool storage up until around 8TB daily
ingest (pre-dedup), above that I would pick 10K disks and a lot of them.

Look for the IBM Linux Tivoli Storage Manager blueprint document for some
additional sizing information in regards to disk layout, but I would highly
recommend using SSD for any TSM server sizing that use dedup (with
replication it's really make a huge difference).





On Sun, Oct 5, 2014 at 12:04 PM, Robert Ouzen <rouzen AT univ.haifa.ac DOT il>
wrote:

> Hi to all
>
> I am in phase of running TSM for Replication for my two backup servers,
> both of them in version 7.1.1 with O.S Windows2008R2 64B  E5530 @ 2.40GHz
> (2 Processors)
>
> When I tested it,  my replicate server was to a Windows2008R2 64bB with
> 64GB memory.
>
> Now for production I am wondering what will be the correct environment to
> do it , your input guys who are already doing it will be really appreciate.
>
> O.S using ? How many core ? Memory ? etc etc ….
>
> Only critical nodes at stage one will be replicate using different
> policies as version 7.1.1 allow it.
>
> T.I.A and best regards
>
> Robert
>