Re: Powering off hosting partition -- MIDRANGE-L

Of course this is not new news that IBM i has this support. I can recall several situations where customers did 'silly things' like pulling the power cord from the right side of an 810 thinking it was redundant. Oops. But plugging it back in the disks spun again and all was well in a few minutes.

At the annual COMMON conference we tripped a breaker that dumped the power to an 0595. That had 1/3 of the disks for the host partition. The host and all guests went A600 until we un-blew the breaker and let it all spin back up. No harm no foul.

IBM i is Amazing!

As Roberto points out though during A600 there am no programs running. The 'suspended animation' occurs for everything including programs that don't need any access to disk such as your UPS program. While I suppose it would be nice for that program to keep running, what could it do? It can't do PWRDWNSYS because it can't get to disk!

- Larry "DrFranken" Bolhuis

www.frankeni.com
www.iDevCloud.com
www.iInTheCloud.com

On 11/5/2014 2:23 PM, Roberto José Etcheverry Romero wrote:

Rob,

If i understand correctly, you tried to have a program run in a hosted
partition while the DISKs for said partitions were unavailable due to the
hosting partition being offline? In any other OS (AIX included) that would
have resulted in a crash, kernel panic or guru meditation (whatever the
cool kids are calling it nowadays). AFAIK only IBM i's LIC has the ability
to simply halt all processing until the disks re-appear. But it's just
that, all processing HALTS (It's one of the very good reasons for a Dual
VIOS setup).

Back on topic, were you trying to see how much time elapsed between the UPS
notifying the hosting partition and the system going down? or was it more
of a "i want the hosting partition to notify all others to come down before
i do"?

On Wed, Nov 5, 2014 at 4:16 PM, <rob@xxxxxxxxx> wrote:

Hey Larry,
I tried that powering off the hosting partition and see what happens to
the guests.
Yes, the HMC shows a SRC that indicates hung disk on the guests.
Starting the host and the guests carry right on.
Instead of power cycling the host a simple vary off of the nwsd should
work.

All happened because of some bad system values while doing a UPS test.
Luckily it was the 'backup' box.

What it did show me was the trying a 10 second loop counter while waiting
for power back up was not the greatest idea. Basically I was seeing if it
was just a temporary power outage that our UPS could easily handle. The
loop didn't increment during the disk hang. Changed the logic to do a
RTVSYSVAL QDATETIME at time of CPF1816 and periodically while waiting for
CPF1817 or a time interval. And calculating the difference between the
two to see if the power didn't come back up in a timely fashion.

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

--
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.