RE: H/A Replication Vendors and Temporal Tables

Well PowerHA is an IBM product so IBM is really the vendor. It is
replication at the storage level, not at the application level.

PowerHA works with internal storage as well as SAN (or both). The set up
you describe below certainly will work but not what we have in mind.

I always use GeoMirror in Async mode. That allows the systems to be a fair
distance apart (More than 50Km)

With SAN storage you can use Flashcopy which will allow you to do full saves
without interrupting the production or replicated server (very momentary
quiesce) So make the flash copy, bring that version up back it up, kill it
and wait for the next one.

If you want to do backups from the replicated system with internal storage,
simply stop replication, do your backup and restart it. Clearly it will
take some time to resynchronize the data in that case. (with Flashcopy
there is no replication delay)

PTFs: do the target, then switch, do the primary, switch back. There is
some minor downtime with the switch if your truly paranoid (I am), but it's
short.

Downside: Some claim that PowerHA uses more communications bandwidth than
the application layer vendors, and I tend to agree, it does use a bit more
bandwidth. It also should have at least a full core on the partitions as
well, so it uses a bit more horsepower than the other guys do. In the end
it's do you want to spend lots of time managing the replication or give it
more machine.

--
Jim Oberholtzer
Agile Technology Architects


-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Rob
Berendt
Sent: Friday, September 09, 2016 7:24 AM
To: Midrange Systems Technical Discussion
Subject: RE: H/A Replication Vendors and Temporal Tables

I wasn't aware they were a replication vendor.
Jim, I know that we've talked about my possible misconceptions before.
When I think about PowerHA I think it's more for hardware replication. You
store your data on a SAN and share it between two boxes. If you lose your
backplane or something else which rarely ever blows you fail over to the
other and keep on running.
I'm more concerned about putting PTF's, OS upgrade, etc on the backup, bless
that. Switch over to the backup, upgrade the primary, switch back.
We also do our daily backups from the backup and the primary only gets
backed up once a quarter.

We're currently not using SAN for our Power 8 boxes. Internal disk hosted
by an lpar of IBM i. Guest lpars are IBM i and AIX.


Rob Berendt