On 07/05/2018 10:14 AM, R.S. wrote: > W dniu 2018-07-05 o 09:36, Jake Anderson pisze: >> Hi >> >> I understand that the operating system timing takes from HMC. >> >> Still we find a few seconds of difference in Operating system. >> >> Does it takes time to percolate and get in sync ? > > Your assumption is not true. > If you don't have STP, your time is just set by operator command. For > STP enablement *and* NTP connection you time is set (during POR) by > SE/HMC and synchronized with external time source. > Perhaps things have changed in the last several years, but it used to be that without an STP the Processor TOD clock was still set from the HMC clock at a POR, and the LPAR "virtual" TOD clock was then initialized from the processor TOD clock when the LPAR was activated. Depending on CLOCKnn PARMLIB member, the Operator might have an option to adjust the clock at IPL and that would also change the LPAR TOD clock. Changing the local clock time after z/OS is initialized only changes the offset of local time from the LPAR TOD clock, not the LPAR TOD clock itself.
In practice, we always reset the HMC clock as close to real time as possible before a planned POR and never needed to fine-tune the clock value from the z/OS console unless we were doing a DR exercise for a different date/time than real time. All LPAR TOD clocks increment in sync, but if two LPARs on the same processor had their LPAR TOD clocks set manually to slightly different times at IPL rather than being controlled by an STP, they would continue to have the same offset indefinitely. The dynamic correction of TOD times only occurs when an STP is in control of the LPAR TOD clocks. Changes to the HMC clock after POR and LPAR activation have no effect on LPAR TOD values.. Joel C. Ewing -- Joel C. Ewing, Bentonville, AR [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
