On a larger system, or if the system has a lot of non critical path files,
(spoolfiles ,test accounts, source code) I would recommend to split the
restore into three parts.

1. restore @[log in to unmask];SHOW;OLDDATE;KEEP;CREATE  (this gives access to all the
utilities)
2. REBOOT and start the network (I like to see if the reboot will work ASAP)
3. restore @[log in to unmask] ACCOUNTS;SHOW;OLDDATE;KEEP;CREATE
4. checkout production applications
5. restore @[log in to unmask]@;SHOW;OLDDATE;KEEP;CREATE

This method while slightly more complex has three benefits:
a. you get to the reboot step sooner. While this doesn't save time it can be
disheartening to find out after 6 hours that you don't have the correct
configuration.
b. you get to the testing and production sooner.
c. the busier files are reloaded together and this creates better locality
on the discs.

ONE OTHER THING,
AFTER EACH RESTORE IN THE SYS ACCOUNT WITH THE CREATE OPTION, IT IS VERY
IMPORTANT TO CHECK FOR USERS THAT DON'T HAVE PASSWORDS! Users in the SYS
account can become group users for PUB and other groups in the SYS account.

This has long been a bug in MPE RESTORE. <rant on> Why the creator of a file
must exist to restore the file when it didn't exist when the file was stored
is beyond me. <pun on> But it does CREATE havoc at restore time!<pun off>
<rant off>

Mike Hornsby
Beechglen Development Inc.
Enterprise Solutions & Support
www.beechglen.com
513-922-0509