Re: SAVLICPGM changing QBRMS?

Hi and thanks for responding.

We're using iTera, replicating to IBM Sterling Forest and they told me to run a
manual savlicpgm so they can restore them to SF.

DR Testing is fairly ez with BRMS and I just had a clean one about 6 months
ago.

2 issues I have;

1. Why/how did QBRMS get changed (the change was password *none as displayed
via dspaudjrne) during savlicpgm.
2. Why did the complete savlicpgm take 3 hours?





-----Original Message-----
From: joberholtzer@xxxxxxxxx
To: midrange-l@xxxxxxxxxxxx
Sent: Sat, 7 Apr 2007 9:22 AM
Subject: Re: SAVLICPGM changing QBRMS?


<snip>
I'm on an 825 w/V5R3, and just ran a savlicpgm and selected all licpgm's
and it took 3 hours to complete. I saved to a 3590E.
</snip>

There are a combination of items being saved and then formated into a
product load as John has already described. BRMS seems to be at the
heart of your question. BRMS has several components that are included
in the *IBM save type, and an *ALLUSR type. That is to say that when
an *ALLUSR save is done, the library QUSRBRM is saved because that is
where all of the databases that allow BRMS to manage the save/restore
environment are kept. Similarly the job schedulers (both advanced and
basic), and other LPPs have components that are part of a *ALLUSR type
save. There are also components of each in the IFS. (John pointed all
this out as well) I suspect you saved a great deal more data than you
thought you were going to since the goal of the SAVLICPGM is to be able
to run the RSTLICPGM which includes all the above mentioned items.

The real question that I have to ask is, what is the recovery plan for
this box? BRMS allows you to create a recovery/save plan that is
tailored to your environment. In a save while active state, BRMS backed
up nearly 8TB of data to and LTO-3 (i595 with 12 processors in that
partition and a TS3310 library with Fibre attach) in about 5.5 hours.
This included everything but QSYS. Yes the tape is faster, and it is an
i595 but your 825 would be slower only due to the tape IOP/IOA/Drive
speed. Instead of doing individual saves, set up the control group to
1) *SAVSECDTA
2) *SAVCFG
3) *IBM
4) *ALLUSR
5) *LINK
6) *ALLDLO

Set BRMS to do a save while active, pushing all the checkpoint messages
to a common message queue, (I use BACKUP) and let it rip. I generally
use a full save on Saturday or Sunday and incremental saves during the
week. BRMS then will print a daily recovery report to show you how to
recover on any given day. All of the user data is on tape at that
point, recoverability is ensured, and its about 90% hands off. You may
have to exclude the performance data libraries and the management
central objects that are a bit of a pain from the save but those objects
can be recovered separately and are recreated on a reload so they are
not required for a recovery. If you run any flavor of Domino or WAS
you will also have some log files to exclude but again they are not
required for a recovery.

Bottom line, I wonder if you are working a bit too hard considering the
BRMS will do most of this for you while you watch. HelpSystems ROBOT is
capable of doing all of these things as well, but you mentioned QBRMS in
the original note.