We're getting started with IBM's Full-System Flash Copy product, which
clones a partition in our V7000 that we can use for backups, "what-if"
games, or other projects.
A controlling LPAR talks to the HMC, V7000, and the live partition and
tells them what to do to clone the live partition into a (thinly
provisioned) Consistency Group. The Consistency Group looks like the
live partition, but is totally separate from the original.
When the process starts, the live partition is Quiesced (forcing all data
to disk), a "table" is created to track which pieces of the live
environment have been copied to the clone partition, and an IPL of the
cloned copy partition is started.
We want to automate the whole process-- start the clone of the production
LPAR, IPL the clone, run an Option 21 backup, and issue the command to
shut down the clone.
A single command gets us through the start of the IPL of the clone.
Since we can talk LPAR-to-LPAR to the controlling LPAR, I think we'll be
able to automate that part of the process.
There are 2 parts of the process that we need to automate:
a) Deallocating memory from one LPAR and allocating it to the clone LPAR
(once it's awake enough to talk to). We've found enough information that
I think will let this happen.
b) What appears to be a sticking point is the operations on the cloned
partition. This is an IPL after an abnormal termination. Any terminated
jobs are ended during the IPL, each of which generates a QSYSOPR message.
At that point in the IPL process, there has been no chance to chance to
change QSYSOPR to *NOTIFY mode, so we get messages for several hundred
spool writers that ended abnormally. Each of which requires a press of
the ENTER key. There are also 3 screens in the IPL process (select IPL;
IPL Options; Access Paths to Rebuild) that require some key presses as
well. There's also a signon screen somewhere in there.
Then there are the commands to issue to run an option 21 backup. I think
we can copy IBM's program; we'd 'only' need to start that program to run
the backup (hopefully without any messages to deal with!).
We're looking for a method that will let an external process monitor the
LPAR being IPLd, and respond to the messages. This all happens before
we're even in restricted state. Since this is controlled by the HMC,
there's no PC-based console to run a program on. Yes, the HMC is a PC...
running LINUX... so any solution I conceive of would have to run on the
HMC under LINUX. Maybe commands from a CL program in one LPAR talking to
the HMC to monitor a green-screen session of the cloned LPAR and respond
to messages there...
I hope we're not the first to the party!
Any hints or nudges in the right direction will be cheerfully accepted!
--Paul E Musselman
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
This mailing list archive is Copyright 1997-2014 by MIDRANGE dot COM and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available here. If you have questions about this, please contact