ADSM-L

Re: [ADSM-L] TSM server scaling/sizing for lots (>20000) nodes

2008-04-30 09:49:36
Subject: Re: [ADSM-L] TSM server scaling/sizing for lots (>20000) nodes
From: Thomas Denier <Thomas.Denier AT JEFFERSONHOSPITAL DOT ORG>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Wed, 30 Apr 2008 09:48:23 -0400
-----Remco Post wrote: -----

[On calculating the average size of a TSM database entry]

>select used_pages * page_size from db
>select sum(num_files) from occupancy
>
>divide the first by the second number and multiply by the number of
>copies you keep for each file
>
>In my home-environment: 806 (very small database, very few files)

Directories, special files, and zero-length files often have
database entries but no data in storage pools; TSM stores all the
data needed to recreate such objects in the TSM database. When I
execute 'delete filespace' commands the number of objects deleted
is typically about 10 percent greater than the number of objects
indicated by storage pool occupancy queries. If the same ratio
applies to your workload, your calculation is overestimating the
average size of an entry by about 10 percent.