Re: System Backup Strategy -- MIDRANGE-L

According to
https://publib.boulder.ibm.com/infocenter/iseries/v5r4/index.jsp?topic=%2Fcl%2Fsavchgobj.htm
:
"To determine the date and time that an object was changed, run the Display
Object Description (DSPOBJD) command with DETAIL(*FULL) specified. For
database file members that were changed, run the Display File Description
(DSPFD) command"

So the date it uses is the object date, not the library date.

And Rob's description of the restore process was the idea behind it.

On Wed, Oct 2, 2013 at 9:17 AM, Luis Rodriguez <luisro58@xxxxxxxxx> wrote:

Rob,

The default value for SAVCHGOBJ uses, as a reference date, the last
SAVLIBdate. So, for example, if you do a
SAVLIB on Sunday and SAVCHGOBJs the rest of the week, your Friday save has
all your changed objects since Sunday.

The date in UPDHST() is used (AFAIK) as a reference date for APYJRNCHG. So,
if you are using journals, restore your Friday and Monday backups and then
do an APYJRNCHG, UPDHST(*NO) would assure a really nice mess (I think).

Regards,

Luis Rodriguez
IBM Certified Systems Expert — eServer i5 iSeries
--

On Wed, Oct 2, 2013 at 9:23 AM, <rob@xxxxxxxxx> wrote:

Perhaps so instead of multiple restores after his full system restore he
can just do two restores
- full restore
- changed objects

A cumulative changed objects versus multiple incremental changed objects.

But I don't use save changed objects so I may not know jack.

Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1
Group Dekko
Dept 1600
Mail to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com

From: Luis Rodriguez <luisro58@xxxxxxxxx>
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>,
Date: 10/02/2013 09:50 AM
Subject: Re: System Backup Strategy
Sent by: midrange-l-bounces@xxxxxxxxxxxx

John,

Just curious, Why did you use UPDHST(*NO)?

Regards,

Luis

Luis Rodriguez
IBM Certified Systems Expert — eServer i5 iSeries
--

On Wed, Oct 2, 2013 at 9:10 AM, John Jones <chianime@xxxxxxxxx> wrote:

One thought would be to ENDTCPIFC instead of ENDTCP. Then you could

give

the system an additional IP address that only the operator's

workstation

would use (if need be, you can config TCP and/or your network to only

allow

the operator station access on that IP).

Example:
Current system uses 10.10.1.5 for all connectivity.
Add a TCP interface of 10.10.1.254 and configure the operator

workstation

to use it.
Update the program to ENDTCPIFC on 10.10.1.5 but leave 10.10.1.254

alone.

Remove the ENDTCP.

This isn't a perfect solution but it may be good enough.

Also, you can very likely reduce your 5 minute delays to more like 1

minute

or even less unless you've some specific job that can take a long time

to

shut down on the new hardware.

To Rob's comments about IPLing, a long time ago I moved to monthly IPLs

and

never looked back. Daily differential backups were save changed

objects

with UPDHST(*NO). The monthly pre-IPL backup was a full system save.

So

any recovery was simply 1) restore the latest monthly followed by 2)
restore the latest daily. Keep it simple.

Even then, monthly IPLs were more a matter of practice than of

necessity.

It also meant that I could easily d/l PTFs any time during the month

and

do

a delayed apply. The next IPL would apply them. (Just don't forget a
change control record if you track changes)

On Wed, Oct 2, 2013 at 6:43 AM, <rob@xxxxxxxxx> wrote:

That last line throws a wrench into a bunch of ideas, but I suppose

you

have your reasons. <snip>FWIW, BRMS is not an option.</snip>

Because

BRMS does have a lot of bail out points, etc.

I know IBM had something called a thin console at one time. But the
documentation I've read said that it's not supported on Power 6 or

Power

7

machines.

http://www-947.ibm.com/systems/support/i/planning/upgrade/futurehdwr.html

Lan Console, HMC or something else I am hearing about FSM - Flex

Systems

Manager has some potential. Verify what you really have. Because if

you

do have a FSM then there are ways to start your system console

sessions

on

it, and get remote access to them that's even better than a PC

Anywhere

type solution. I know for a fact that HMC has this. Do you have the
model number? Can you snap a picture of the screen and shoot it to

me

offlist?

Someone mentioned a point about your weekly IPLs. There are

pros/cons

to

IPLing.
If you have an automated routine that downloads the latest cume and

PTF

groups and sets those to apply every week then a frequent IPL will

keep

you up to snuff on patches. Prompt SNDPTFORD and look for the

special

values like *CUMPKG, *HIPERGRP, etc.
If you use SQL, imbedded in your applications or elsewhere, IPLs

destroy

some things that help SQL to run faster. Let me explain it like

this,

you

run an sql statement. It analyses it and figures out the best way to

run

it, and runs it. The next time you run it it checks a special area

and

knows now how to run it and skips that step. You IPL and that is

cleared

and it has to reanalyze it.
Back when disk drives were measured in MB instead of GB the amount of
space cleared up by an IPL was quite significant. It just isn't so
anymore.
If you frequently IPL you'll learn quickly to add stuff to the

startup

program. If you only IPL every three years you may forget how to

start

up

all those WAS based applications and whatnot.

Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1
Group Dekko
Dept 1600
Mail to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com

From: "Buzz Fenner" <bfenner@xxxxxxxxxxxxxxxx>
To: "Midrange-L Discussion" <midrange-l@xxxxxxxxxxxx>,
Date: 10/01/2013 03:56 PM
Subject: System Backup Strategy
Sent by: midrange-l-bounces@xxxxxxxxxxxx

We do production backups every business day, and save the entire

system

plus
IPL the system on the last day of the week.

We don't use the GO SAVE menu to accomplish this 2nd backup; instead

it's

an
in-house-written program equipped with prompts (and bailout points)

for

the
user and tape volume validation code. It's also interactive up to the
point
of issuing the power down with restart. The user acknowledges a

series

of

prompts along the way. This job runs in QCTL. (We don't have real
properly-trained system operators to handle these tasks, so the boss

did

not
want to give them that sort of system access. These are part-time
employees
and I see his point.)

We changed over from a Power 5 to a Power 7 last weekend, so no more
twinax! (Yes, we had twinax devices.)

We had modified our code (with what also had been fixed via ptf, that

is)

to
get TCP shutdown before ending the subsystems. It's a combination of
EndTcpPtp, EndHostSvr, & EndTcp with 5 minute delays sprinkled in

between

them.

Now the twinax operator's workstation has been replaced with a thin
client.
The program as written no longer works because communication is

terminated

with the workstation before they have the opportunity to answer some
prompts.

A guy with our Business Partner suggested loading up the Ops Console

from

iSeries Access For Windows and somehow giving the operator a back

door

into
the system.

I'm thinking of rewriting the process that does all the input

validation

and
gives the operator the option of submitting a batch job (to QCTL) to
complete the task.

Better ideas, suggestions, or am I just way off base here?

FWIW, BRMS is not an option.

--

Buzz Fenner

Business Systems Analyst/Systems Administrator

City Water & Light

870.930.3374 | 870.219.5229

bfenner@xxxxxxxxxxxxxxxx

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)

mailing

list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L)

mailing

list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

--
John Jones, CISSP
No security, no privacy. Know security, know privacy.
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing

list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing

list

To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.