>Is there any truth to the rumor I just heard about performance degradation
>when the ADSM data base reaches 40GB?

John - I haven't heard talk of that specifically, but there have been
       general discussions about feasible server sizes on the List
(maybe a year or so ago).  The consensus is that when your server
reaches "huge" size that you subdivide into multiple servers.
ADSMv3 facilities particularly lend themselves to that.  The various
discussions on the List regarding performance have shown where
single-threading limits ADSM throughput and hence poses some limitations
on what can be done, though a nicely configured ADSM server (multiple
storage pools, database spread over a bunch of volumes) can help ADSM
scale pretty well.  The big cloud that hangs over a large server, of
course, is the prospect of damage to the database requiring recovery
of the whole thing.  The ADSM architecture mitigates this by supporting
two ADSM mirrors of the database and recovery log, but there can be
logical disasters of various kinds as well which no mirror will help.
     Richard Sims, BU