Redundantly reporting on multiple LPARs for hardware errors is not bad either.
Many customers including us have true sandbox LPARs where automation may be limited or operators may be conditioned to not pay attention to it. I think for true hardware failures reporting on every applicable LPAR is a good thing. It makes it likely that one of those LPARs will include automation prepared to handle the message and that it will be visible to operations staff. It also helps operations to understand this is a hardware problem with processor x when it comes out on n of n LPARs and all the impacted LPARs reside on the same processor. Best Regards, Sam Knutson, GEICO System z Performance and Availability Management mailto:[EMAIL PROTECTED] (office) 301.986.3574 (cell) 301.996.1318 "Think big, act bold, start simple, grow fast..." -----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Jim Mulder Sent: Monday, December 08, 2008 12:45 AM To: IBM-MAIN@bama.ua.edu Subject: Re: z10 power problem notification You are correct, those things are the job of the service processor/HMC. In the "good old days", we did not have a service processor with the capability to perform those functions, so the operating system had to do it. It now makes more sense to use the service processor, so that we do not need to provide those functions in at least 4 operating systems (MVS, VM, VSE, and Linux), and so that multiple LPARs are not all reporting the same condition. We do not like to put model dependent code into the operating system when that can be avoided. > It seems odd to me to go down the path of ill-tested SNMP or emails from the > service processor, when there is a much more robust and *architected* > mechanism already there. Sure, the guest OS can fail because of the hardware > failure, but that'll get noticed PDQ! If you are suggesting that some customers might find it convenient to have the option of having some processor issues redundantly reported via MVS WTOs because they already have an automation infrastructure designed around WTOs, or because the MVS Syslog or Operlog is a convenient place for them to look for such things, or to correlate them with other operating system events, I can't disagree with that (especially since that might be able to work with already existing MVS code, albeit unused for the past decade AFAIK). You may have a reasonable request to make to the service processor folks in Endicott. Jim Mulder z/OS System Test IBM Corp. Poughkeepsie, NY ==================== This email/fax message is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized review, use, disclosure or distribution of this email/fax is prohibited. If you are not the intended recipient, please destroy all paper and electronic copies of the original message. ---------------------------------------------------------------------- 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