Hi,

We have just completed this exercise, replacing 10 3590s with 3590Es
in a 3494.  Try and organise it so that you stop writing, upgrade the
drives and then start writing again.  Otherwise, maybe upgrade one,
introduce a temporary storage hierarchy and get some confidence using
it solely with a few dedicated tapes.  I think it is too complex to
run a mixed production environment of 3590/3590Es.

The microcode you need is D0IC_533.  ADSM server needed is 3.1.2.40.
We have Atape 4.4.0.0; atldd 4.0.1.0.  We have a shared 3494 and one
adsm server is 4.3.2 and one 4.2.1.  All are fine.  The 3494 Library
Manager code needs a specific version - I haven't got a note of that.
The LM and first drive is the only change that needs the 3494 offline;
subsequent drive updates and "teach" can be done on the fly.

Replacing tapes can be a gradual movement.  All filling tapes must be
marked READONLY.  All scratch tapes will automatically be relabelled
and rewritten as 3590E format after the next PENDING/EMPTY cycle.
PRIVATE tapes must be marked READONLY and when they have gone through
PENDING/EMPTY must be set back to READWRITE and then they too get
relabelled completely automatically.  You could make a script to do
this.

FILLING ones should be emtied first with MOVE DATA.  Later you can do
MOVE DATA or reclamation on the rest to gradually move the contents of
existing tapes, as is convenient.

I am waiting to hear about whether the ADSM DB contains information
about 3590/3590E.  We had a curious problem a few weeks ago which I
submitted as a PMR, but have spent several weeks convincing them that
the 3494 *is* working (and not broken for several weeks!); the problem
really was that I believed the ADSM DB wrongly thought 3 3590E private
(and formerly scratch) tapes were in fact 3590s but I can't prove it.
It refused to mount them unless they were READONLY ...  which leads me
to believe there must be a flag in there somewhere but it's not
visible to me!  It was only these 3 tapes though and happened when we
were in validation phase, so hopefully it wouldn't impact on you.

Hope this helps.
Regards, Sheelagh
------------------------------------------------------------------------
Sheelagh Treweek                   Email: sheelagh.treweek AT oucs.ox.ac DOT uk
Sheelagh Treweek                   Email: sheelagh.treweek AT oucs.ox.ac DOT uk
Oxford University Computing Services           Tel:   +44 (0)1865 273205
13 Banbury Road, Oxford, OX2 6NN, UK           Fax:   +44 (0)1865 273275