Greetings, all.


We're about to move the data on our campus mail server (45000 users, ca 60G of
mail) to a new hunk of disk, and are evaluating our filesystem arrangement.

(background data:
ADSM server 3.1.2.1 on AIX 4.2.1
ADSM client 3.1.20.1 on AIX 4.2.1
)
Does anyone have experience or data on optimal sizes of filesystems?
Does anyone have experience or data on optimal sizes of filesystems?
I'm thinking in terms of

Size of database queries when doing normal incrementals
Speed of same for restores
Speed of database work during incrementals

We currently have ca. a million files per filesystem, and ca. 200% turnaround
per week. ( 2 million new files, 2 million deleted files) and normal
incrementals, not -INCBYDATE, can take a very long time, especially if some
systemic failure has kept incrementals from happening for a few
days. (shudder)

Initial incrementals on these filesystems took _days_

Clearly, we'd do better with more, smaller filesystems.  But how much smaller?
Where is the breakpoint?  Are my estimations of bottleneck location way off
base?


Allen S. Rout