On Mon, 28 Apr 2008 10:06:21 -0500, Mark Zelden
<[EMAIL PROTECTED]> wrote:
>Bill, Thanks for the clarification. So why doesn't (can't) SFM prevent
>IXC102A
>after VARY XCF,sysname,OFFLINE if there is an active system in the same
>sysplex on the same CPC? It it just a SMOP, or is there a technical reason.
Mark,
XCF only resets or deactivates an LPAR if told to do so by the SFM policy.
The SFM policy actions do not apply to V XCF commands. They only apply in
status-update-missing scenarios.
And yet, if an SFM policy exists at all, no matter what it specifies, XCF
will
attempt to isolate the outgoing system on a V XCF command. I don't know
why that is. The rationale is not clearly documented in the code, and no one
still associated with XCF development recalls why that implementation was
chosen.
There is work in progress now to design more effective and consistent
mechanisms for removing a system from the sysplex without operator
intervention. It is highly desirable to eliminate operator prompting in this
situation, because a great many outages have occurred over the years when
either (1) the operator or automation replies DOWN prematurely and data base
corruption occurs as I described earlier, or (2) no one ever replies DOWN, and
the sysplex begins to suffer from "sympathy sickness" because resources are
held or protocols of one kind or another are impeded by the unresponsive
system.
Bill Neiman
XCF Development
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
