Enhanced SETIOS Function to Perform Freeze

Thank you for those who have responded. I would like to share our thoughts as we are working on this item. If anyone has additional comments or concerns, please comment here or email me (tpowels@us.ibm.com)

Freeze vs Suspend:

There were some comments in this RFE mentioning CSUSPEND. I would like to be clear that CSUSPEND will not raise Extended Long Busy, and so data consistency on the secondaries cannot be guaranteed following a Freeze event that was triggered due to suspend. As a warning for anyone currently testing with CSUSPEND, I would recommend to consider using CGROUP FREEZE for testing instead.

CGROUP FREEZE will raise the Extended Long Busy condition on the PPRC primary device, as well as suspend PPRC. This allows GDPS or z/OS HyperSwap to perform the consistent suspend.

The new command will issue similar I/O that freezes the LSS pair. In case anything goes wrong, this will similarly provide an option to RUN (thaw) the devices to lower the ELB.

Operator Input for the command:
For now, we are prioritizing the form of the command that takes action on a loaded HyperSwap configuration. In the future we would be happy to consider expanding this to additional use cases like specifying a device range, physical box, CHPID connection, etc.

Keyword naming:
We are thinking to add a generic “HSWAP” keyword to the SETIOS command, with the user specifying the action FREEZE or RUN.

SETIOS HSWAP,FREEZE[,CONFIG=hsconfigname]

SETIOS HSWAP,RUN[,CONFIG=hsconfigname]

By using the HSWAP keyword, we have the flexibility to add additional actions besides FREEZE and RUN in the future.

Regards,

Tabor Powelson, IOS and z/OS HyperSwap Design and Development

Hi

I think this is an excellent idea. From my and my customers point of view, this would be of value
SETIOS PPRC,SUSPEND,LEG=LEG(x)

or pointing to a specific device / device range
SETIOS PPRC,SUSPEND,DEV=device/device range to be suspended

Taking it a bit longer, maybe overkill, but a CHP is a possibility - make a suspend for all devices on CHP(xx)
SETIOS PPRC,SUSPEND,CHP=??

Again, SETIOS freeze would be a very nice enhancement, so a simple V1.0 supporint a single device as a start, and enhance it as time and request comes in, may be a way forward.

Hello all, I am looking into this and would like to get some more details on what is the most useful.
- For those who are interested in this request, how would you prefer that we handle multi-target PPRC? Is there a need to request the suspension of a specific leg of multi-target? Or should the command simply default to the first PPRC leg found?
- Are any users interested in using this for non-HyperSwap configurations, such as a CSM Hardened Freeze configuration?

--
I am considering implementing syntax such as the following for HyperSwap users, which will work for both GDPS and z/OS HyperSwap configurations:

(1) SETIOS PPRC,SUSPEND - Suspend a PPRC pair in the currently active HyperSwap configuration. Note in the case of multi-target PPRC, this will suspend only one HS configuration.

(2) SETIOS PPRC,SUSPEND,HSCONFIG=myHSConfig - Suspend a PPRC pair in a specific HyperSwap configuration.

--
Are those two above forms of the command the most useful, or would it also be helpful to allow specific devices to be input:

(3) SETIOS PPRC,SUSPEND,DEV=pppp - Suspend the PPRC pair for primary device pppp

(4) SETIOS PPRC,SUSPEND,DEV=(pppp,ssss) - Suspend the PPRC pair between primary device pppp and secondary ssss.

Due to processing by IBM, this request was reassigned to have the following updated attributes:
Brand - Servers and Systems Software
Product family - z Systems Software
Product - z/OS
Component - BCP_IOS
Operating system - IBM z/OS
Source - COMMON

For recording keeping, the previous attributes were:
Brand - Servers and Systems Software
Product family - z Systems Software
Product - GDPS
Component - GDPS
Operating system - IBM z/OS
Source - COMMON

medium

Medium priority. Would be a useful feature.

We can only support this RFE as Unplanned-Freeze-testing is even more tedious than Unplanned-HyperSwap-testing.

Medium. It would be good to have this command to simply the test for an unplanned FREEZE.

Nice to have!

I like the idea

Will be medium priority when we get to z/OS v2.4

We would consider this a medium priority for our site.

medium. That is a good enhancement to trigger events easily.

This is not a request at the time but maybe we will testing freeze in our disaster test at a later time.

If this was provided we'd probably continue using the CSUSPEND command as it is already very simple and I think a better validation of an example triggering event. We'd box a device manually instead of using SETIOS to trigger unplanned Hyperswaps had IBM not removed support for that strategy.

Yes, SETIOS for HyperSwap is good and we use regularly. Would like equivalent for a freeze, rather than using CSUSPEND. Not urgent though.

this is a very logical extension of the SETIOS command, now that we have the hyperswap option
"medium" importance

this is intended to provide clients a capability to simulate an unplanned PPRC suspend