Re: AS/400 Backup questions (was: Probably very easy, Newbie question ) -- MIDRANGE-L

I wrote

> I have said in prior posts that in recent years there have been a variety of
>  different ways that we do our backups.

The current method might be called GO SAVE 21 UNATTENDED or no operator
involved once it starts good.

The old method was GO SAVE 21 with OPERATOR guiding it from start to finish
... the reason for the old method was mainly our inexperience with the system.

>  This is how we NOW do backups.
>
>  1. Some time before backup to be done ... check out what is running on the
>  system, who is signed on, are they really active.

Step 1 was pretty much the same deal.

<snip>
until we got good with WRKACTJOB variants
we were doing lots of phone calls to remote sites

>  2. Tape drive cleaning (not every nite) & Initialize the tape (I use a menu
>  option setup in advance for this... INZTAP is the command) & plan whatever
> me doing this evening
> so I can see where doing the backup will be a convenient point.

We now use naming convention of IBM01 for Monday nite backup
IBM02 IBM03 IBM04 IBM05 Tues Wed Thu Fri

Because I managed to persuade management that it was more economical to buy
IBM brand that goes several years before it flakes out, as opposed to el
cheapo which dies when you least expect it & then you have to run that backup
again & how do you know that last nite's backup is any good ...IBM log might
say everything went fine, but the media is about to die.

When we were using different brands of tapes from different vendors that sold
tapes & they died before the warranty ran out, we had to know which vendor
sold us this when, so I had a tape naming convention that included 2-3
letters for each of:
Brand name of media
Vendor that sold it to us
What density is this thing ... 2 Gig or whatever
When it arrived in our system
When we first used it
digit or character on end ... tape A B C D etc. where all the above stuff the
same

I had an INZTAP prompt screen CL in which I would fill in name of vendor etc.
to get the naming convention taken care of

Every time after some tapes died, I needed to review both naming conventions
... what shall we call the next tapes, and problem history, so I could fight
the battle to get EL QUALITY instead of EL CHEAP

STRSST /1 /1 /4 activity log on removable media

To initialize tapes if you have not setup what you always do on some menu
like I have

INZTAP F4
or GO TAPE /2

>  If orange light on tape drive
>  that means we have left it too late to do a proper cleaning
>  so lets clean it twice

We did not know that originally ... we thought the OJ light meant time to
clean it once.  This probably contributed to more rapid demise of our tape
media & drives.

>  If green light blinking on tape drive
>  wait until solid before doing next step

We did not know that either ...
which probably contributed to some of the problems we had

>  3. Get all jobs to a close, including printers.

We learned the hard way with GO BACKUP that even though the 400 could backup
the important stuff concurrent with printer running (big help when end-month
got a few thousand pages to print) it was not worth the headaches digging
into error messages after backup to make sure the only reason we had the
error messages was the printer was running.

>  In some cases I am in JOBQ
>  putting stuff on hold & jotting down what I put on hold so I know what to
>  release later.
>
>  Notice that when putting on hold ... I work from the LAST job that will
run,
>  backwards, then after backup, I work from the NEXT job that will run,
>  forwards.
>
>  WRKJOBQ gets you at all the JOBQ in your library list ... WRKJOBQ F4 to be
>  sure of getting at all the ones in the system.
>
>  Same kind of thing with reports.

I think it is WRKUSRJOB *ALL *ALL that gets you all non-system jobs on your
system, including running right now, waiting in JOBQ etc & by futsing around
with F4 you can exclude from consideration those that have gone to a
conclusion.

>  4. Stop All Subsystems.
>
>  There will be people here who say that this is handled by the GO SAVE 21 &
>  you not need to do it twice, and then I talk about jobs that we have
> problems
>  with & they say tell IBM its a bug, well there has been multiple talk about
>  3rd party stuff that violates IBM rules ... we just find that we have NO
>  PROBLEMS when we do it our way & we used to have problems when we did it
>  another way.

I may do a later post pointing at some of the gotchas that we found out
about, and cared enough to document.

>  5. GO SAVE 21

Since we were also doing GO BACKUP a lot, our instructions reminded people
which menu they should be on.

Also gotta be on the main console signed on as Master Security Officer.

We were klutzes so we put ourselves into *BREAK message mode just to be on
the safe side, even though a later step does that for us.

When there are already a lot of unanswered messages on QSYSOPR that are
unrelated to the backup, let's review them before starting the backup.

>  There is the screen that lists all the stuff involved in the backup
>  then a screen to fill in / adjust some blanks & this is what we do
>
>  TAP01 (default)
>  N   Prompt
>  N   Check
>  *NOTIFY because we think we know what we are doing
>  *CURRENT    (default)
>  *ALL servers
>
>  scroll to second screen
>
>  N (default)
>  N (default)
>  Y (reply list)

Take your time filling in the screen & roll back & forth to make sure answers
right

You really really do not want to be doing it the way we used to do it ... it
would stop every 10-15 minutes with another screen for us to fill out & we
had copious notes how to fill out each screen

Quite recently we had this scenario:
The boss told me he was done with end-fiscal stuff & I could do the nitely
backup.
I started the backup.
He came in saying there was one more report he wanted to do & could I
interrupt it.
I told him, yes I can & I want you to see how, because I have never shown you
how to interrupt a dedicated restricted state task (my boss does backups when
I am on vacation), but I warn you in advance that the time lost will be
significant.
How significant he wanted to know
I guess about an hour or so
Ok show me.
So I showed him
We lost like 2 1/2 hours (the backup normally takes 1 1/2 hours)
I told him that now he knows how to do a really useful thing, but the lesson
is that if you need something in the future, it will be done faster if we
wait on the backup getting finished & if you want to leave, perhaps leave me
instructions how to do it when the backup is done.
This was a month before the tape drive died & when that was fixed I learned
some stuff about cleaning tape drive more aggressively to get better
performance.

>  If green light blinking on tape drive
>  wait until solid before doing next step

If anything goes wrong with the backup, such as error messages at the end.

SIGNOFF *LIST so there is a detail job log on the episode.
Then print with SAVE(*YES) first.

MacWheel99@aol.com (Alister Wm Macintyre) (Al Mac)
BPCS 405 CD Manager / Programmer @ Global Wire Technologies Incorporated
http://www.globalwiretechnologies.com = new name same quality wire
engineering company: fax # 812-424-6838