RE: Full system save 21 schedule

Paul--

1) Our two systems are two LPARs on one system. We actually have 8 LPARs in total: 2 production LPARs; 2 Full-System Flash Copy LPARs (one for each production LPAR); a 3rd production LPAR used by 2 people (an ancient app lives there); a Flash Control LPAR; and 2 VIOS LPARs.
3) We have no DASD; everything is stuffed into a V9000 SAN. All flash memory (not to be confused with a Full-System Flash Copy (FSFC)). We have a lot of 'extra' space-- some of the network servers have SANs nearing end-of-life; they're destined to move into the V9000 'in the future.'
2) The two production LPARs are 10 Tbytes allocated. The two FSFC LPARs are 10 Tbytes, "thinly provisioned." Model 8286 42A EPXE. Memory 128 Gig and 123 Gig in the production LPARs. Total memory on the box ~329 Gig.
4) The flash copy takes about 2 minutes-- Quiesce the LPAR, flash the disk (well, storage), release the LPAR.
5a) Mostly home-grown order entry/processing; accounting.
5b) If you're talking about the backup, it's a home-grown process that starts on the LPAR to be backed up, pokes at the controlling LPAR to flash the system, then home-grown backup process on the flashed copy. We're using Batch Restricted State for the backup, and running an Option/21 backup every night. We give the Batch Restricted state a time limit of 16 hours, so we know it will wake up eventually!
6) BRMS might, but we've always been a ROBOT/Save shop (long before BRMS was a gleam in IBM's eye!). When we started configuring the backup it seemed like less work to DIY, rather than configure BRMS. This may not have been true! (; The process started as a fully-manual process, with pieces automated as they became annoying (allocate a tape drive; determine which tapes are available for the backup; save logs &c; send the logs &c back to a production LPAR for safe keeping)

The backup takes about 13 hours, including the setup, running the backup, and saving logs &c from the flash LPAR back to one of our production LPARs. It may seem like a long time, but we have the resources available. Plus, we end up with 2 tapes that have our entire production system.

Paul E Musselman
PaulMmn@xxxxxxxxxxxxx

-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Steinmetz, Paul
Sent: Friday, March 23, 2018 8:50 AM
To: 'Midrange Systems Technical Discussion' <midrange-l@xxxxxxxxxxxx>
Subject: RE: Full system save 21 schedule

Paul,

1) Are your two systems two lpars on one system or 1 lpar each on two systems?
2) How big are your lpars(systems), CPU, MEM, DASD?
3) What type of DASD?
4) How long does the flash copy take?
5) What apps are you running?
6) Would BRMS fit/work in your solution?

Paul

-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Musselman, Paul
Sent: Thursday, March 22, 2018 2:14 PM
To: Midrange Systems Technical Discussion
Subject: RE: Full system save 21 schedule

We're fortunate that we have a v9000 and enough empty space to do a Full System Flash Copy ever night, and run an Option/21 daily.

Before the V9000 we had a V7000 with a TMS-820 'flash appliance, and that's when we started the daily Option/21. Prior to that, the only time we were able to sneak in an Option/21 was every 3 years in anticipation of either a System or Hardware upgrade!

Before we had the flash capability, we ran an Option/22 every month (OS and IBM libraries and IFS). That, and the daily backups we ran, were our recovery. We have 2 systems, and use MIMIX to copy our production data to our development system. We backed up from the development copy when possible.

We retain our month end backups 'forever.' My goal has been to make every month end backup a self-contained snapshot of our systems as they existed at that point in time. Given a backup, and a large enough machine, we could re-create our system as it existed at that point in time: OS, libraries, programs, and data, all in sync.

Paul E Musselman
PaulMmn@xxxxxxxxxxxxx

-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Jeff Crosby
Sent: Thursday, March 22, 2018 12:59 PM
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>
Subject: Full system save 21 schedule

All,

What do you do/suggest as a full system save schedule?

We did a full BRMS save 5 nights a week with our inhouse system. Now that we are in a data center, they do Evault backups to an appliance.

They will do a full Save 21 on whatever schedule we want. They say most do once a year, but that doesn't seem like enough for me. I'm leaning toward once every 13 weeks, meaning 4 times per year. And I would try to coincide PTF applications to these Save 21s.

Thanks.