-----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Paul Gilmartin Sent: Friday, May 18, 2007 4:29 PM To: [email protected] Subject: Re: Synchronize Time Between Mainframe and Servers?
On Fri, 18 May 2007 16:06:12 -0500, Field, Alan C. wrote: > >You don't need a Sysplex timer to do this. > >The sysplex timer ensures the mainframe(s)have a consistent time. With >or without a timer you can run an SNTP task on the mainframe (its part >of TCP/IP since about z/OS 1.6), otherwise look for Keith Clapp's >excellent SNTP product. > Without a time reference, the mainframe TOD clock will have a secular drift. Will SNTP adjust the TOD clock to the possibly more accurate time elsewhere in the enterprise, or merely adjust the remainder of the enterprise to the incorrect time of the mainframe? <SNIP> 1) Let us assume that the time on the mainframe has been set to within 1 second of a time reference standard. 2) Let us assume that all other processors get the same treatment by the same highly accurate operator. Ok, take a look at a Principles manual (the following quote is from p. 4-27 for the ESA/390 manual -- Chapter 4 Control): "The TOD clock nominally is incremented by adding a one in bit position 51 every microsecond. In models having a higher or lower resolution, a different bit position is incremented at such a frequency that the rate of advancing the clock is the same as if a one were added in bit position 51 every microsecond. The resolution of the TOD clock is such that the incrementing rate is comparable to the instruction-execution rate of the model." I have yet to see any Intel (or other) based system have a clock that is this stable, that the drift is measured in seconds a year, not seconds a month. And this is why I have to use an external reference to dither the clocks on my systems on my LAN, and I think my employer does the same to sync the corporate LAN. So if you have a political problem, how about turning it into a technical issue? If you show them the doc on the cycling of a z/ARCH machine (use the page from a Principles of Operation's Chapter 4, Control) on TOD cycling, and then have them show you the equivalent from their systems. The reason for an ETR is to make sure that all LPARs (on the same CEC or not) in a SYSPLEX are using the same basic time reference, because the granularity of an operator issuing time commands is not that good! But then, my experience with TOD drift against a known standard has been rather remarkable. Quite seriously, it has been only a few seconds over a year's period of time. Regards, Steve Thompson -- STD Disclaimer: my posting may give an opinion that may not be in line with the company's opinion, or that of their attorneys. -- ---------------------------------------------------------------------- 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
