Re: Synchronize Time Between Mainframe and Servers?
On Mon, 21 May 2007 13:08:55 -0500, Dave Kopischke wrote: Also, if you lose the timer, you can recover and run without it. It takes a few seconds for error recovery and to present the option on the console. By this time, most of our communication sessions time out, so we've got some work to do to get everything reconnected. But batch JOBs continue to run and everything else runs just fine too once you restart all your VTAM applications and such. We just had our first opportunity to disconnect the timer and run without it since upgrading to z/OS 1.7. It was a non-event. No stoppage at all. Just a nasty message on the console about losing ETR ports and losing the ETR, but everything kept running. No prompt to answer or anything. And it recovered automatically when we reattached the timer as well. When we were on OS/390 2.10 (and z/OS 1.4 if I recall correctly), it would lock up every LPAR and make you answer a console prompt before proceeding. Lots of interactive things used to time out. This is much nicer. -- 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
Re: Synchronize Time Between Mainframe and Servers?
Today, STP uses the same modem that the HMC uses to download patches from IBM. When STP uses the modem, it simply dials NIST in Boulder, or whatever dial up time server a Sysplex Timer would use in your geography. As often mentioned, apparently IBM plans to enhance STP to be able to synchronize with a NTP hierarchy in the future. Brian On Fri, 25 May 2007 09:37:27 +1000, Shane Ginnane wrote: Glenn wrote on 25/05/2007 08:16:16 AM: Since the ETS call out function of the STP feature uses the HMC you should already have the modem/phone line available ( does anyone run all their HMCs without a modem/phone line, just curious ). Yeah, I thought about this after I posted - the modems we had to find (with RS-232-D plug) were for the timers. Hopefully for STP, there will be sufficient smarts for the current modems to do both functions. Shane ... -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Thu, 24 May 2007 17:16:16 -0500, Glenn Miller wrote: According to the Redbook I mentioned, each STP-capable CEC and CF must have the STP feature code 1021 installed. This feature code is a chargable feature. Your local IBM office would be able to provide you with a price quote. Thanks. And: From: McKown, John Date: Thu, 24 May 2007 08:25:07 -0500 What's so hard ???. Shane ... Well, in our environment, what is difficult is convincing management to spend a large amount of money for a piece of hardware (the ETR) in order to do what other systems get for free. More ammo that the System z is just too expensive. Given the laissez faire attitude towards time in systems such as UNIX and Windows, it is difficult if not impossible to explain why the z/OS TOD clock can't be changed ad hoc while the system runs. It works in UNIX and Windows! What is with that stupid z/OS system? It's IBM's business decision to make the STP feature chargable. IBM has gone the opposite direction with, for example, the hardware support for ICSF. Of course if, in a future model, IBM chooses not to make STP separately chargable, IBM will still need to recoup its expenses by including STP in the total end user price. It's a matter of management psychology: do the purchasing decision makers find it more acceptable to buy a single item, a computer with a controllable clock, than to pay the same total price with a conspicuous line item for the clock that no competitor itemizes? This shouldn't affect the decision of those sites (a majority?) that would purchase STP regardless; only those that face the choice of letting their clocks drift or paying extra. Will there continue to be interface support for (relatively cheap?) aftermarket ETRs? I agree with the design decisions prohibiting changing the TOD clock ad hoc, but I would like to see a larger slew rate, hours per day rather than seconds per day for quick correction of gross blunders, and a larger maximum adjustment, hours or days rather than a minute, both with operator confirmation. Any site that starts up a system with the wrong month or year might as well re-IPL. -- gil -- 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
Re: Synchronize Time Between Mainframe and Servers?
Yes, we're all converging on the correct answers. Here's a quick summary. There are two supported paths for IBM mainframes to synchronize with an external clock source, and then a few different types of external clocks that can be used. The older option is to install a Sysplex Timer (9037-002) and then connect the Sysplex Timer to an external time reference (ETR) via the 9037 serial port. You can use a dial-up modem (to the U.S. NIST modem pool, for example), a shortwave radio receiver (to NIST's WWV for example), or a GPS receiver. GPS receivers are certainly more in vogue now, and there are several companies building such boxes. Configure everything correctly (including operating systems) and you're all set. Note that some people have figured out how to use that serial port connection with other devices that could deliver the time information bytes. Not recommended. Support for the 9037-002 will end at some point in the future, so nowadays IBM is recommending STP (Server Time Protocol) which is supported on z890, z990, z9 BC, and z9 EC systems. At present your external time reference would be via dial-up modem to NIST and similar time source modem pools -- that's supported much like the Call Home feature and easy to set up on the HMC. IBM has a Statement of Direction expressing the company's intention to support two more external time sources in the future: GPS and any arbitrary NTP server (NTPD). My hunch is the latter option will come with all sorts of warning statements -- fall victim to time of day spoofing and you could have horrible problems -- but let's leave that aside for now. :-) At present you have two choices for keeping mainframe time in sync with other servers' time clocks: 1. Make the mainframe boss. This is the best choice for a lot of reasons. Set up the mainframe to get its external time reference, then use the built-in time server so that other servers can get time sync over the network, perhaps in a hierarchy. 2. Have both the mainframe and your organization's master time server (if they are unfortunately different entitites) both sync to the same external time reference (such as NIST dial-up). This might not offer as much accuracy, but it might be sufficient for certain purposes. - - - - - Timothy Sipples IBM Consulting Enterprise Software Architect Specializing in Software Architectures Related to System z Based in Tokyo, Serving IBM Japan and IBM Asia-Pacific E-Mail: [EMAIL PROTECTED] -- 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
Re: Synchronize Time Between Mainframe and Servers?
Timothy Sipples wrote: Yes, we're all converging on the correct answers. Here's a quick summary. [...] At present you have two choices for keeping mainframe time in sync with other servers' time clocks: 1. Make the mainframe boss. This is the best choice for a lot of reasons. Set up the mainframe to get its external time reference, then use the built-in time server so that other servers can get time sync over the network, perhaps in a hierarchy. That means buy 9037 or STP (it's not free) and connect it to every CPC you have. I don't know STP prices, but 9037 costed approx. 50k$. You need at least two for redundancy. If you cannot connect all CPCs to the same set of timers, you need more timers. For STP - I'm not sure, but the price is per CPC. Obviously multi-CPC sysplex users already have one of theses solutions. So, (before STP was GA), the price for time reference in mainframe is 100k$ *more* than any other server. Piece of cake. Anytime I provide the cost my management say F*ck, mainframe will remain out of time synchronization BTW: SNA connection does not allow to use mainframe as time source. I heard about installations, where IP network is strictly isolated from mainframe (although it has its own IP network). Of course this is not a problem with mainframe itself. I can be a problem of any machine isolated by someone's decision. -- Radoslaw Skorupka Lodz, Poland -- BRE Bank SA ul. Senatorska 18 00-950 Warszawa www.brebank.pl Sd Rejonowy dla m. st. Warszawy XII Wydzia Gospodarczy Krajowego Rejestru Sdowego, nr rejestru przedsibiorców KRS 025237 NIP: 526-021-50-88 Wedug stanu na dzie 01.01.2007 r. kapita zakadowy BRE Banku SA (w caoci opacony) wynosi 118.064.140 z. W zwizku z realizacj warunkowego podwyszenia kapitau zakadowego, na podstawie uchwa XVI WZ z dnia 21.05.2003 r., kapita zakadowy BRE Banku SA moe ulec podwyszeniu do kwoty 118.760.528 z. Akcje w podwyszonym kapitale zakadowym bd w caoci opacone. -- 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
Re: Synchronize Time Between Mainframe and Servers?
-Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Shane Ginnane Sent: Wednesday, May 23, 2007 6:34 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? I can't believe this thread is still alive. Everyone (presumably) has an in-house time server for their network - possibly excluding the mainframe. Go connect your mainframe (ETR or STP) to an atomic source and let it stabilize (if you don't have either, you'd better start looking at the latter). Then either use *it* as your in-house NTP time-server, or put it as the first server on the list for the box that currently dials out for the time. What's so hard ???. Shane ... Well, in our environment, what is difficult is convincing management to spend a large amount of money for a piece of hardware (the ETR) in order to do what other systems get for free. More ammo that the System z is just too expensive. Given the laissez faire attitude towards time in systems such as UNIX and Windows, it is difficult if not impossible to explain why the z/OS TOD clock can't be changed ad hoc while the system runs. It works in UNIX and Windows! What is with that stupid z/OS system? -- John McKown Senior Systems Programmer HealthMarkets Keeping the Promise of Affordable Coverage Administrative Services Group Information Technology The information contained in this e-mail message may be privileged and/or confidential. It is for intended addressee(s) only. If you are not the intended recipient, you are hereby notified that any disclosure, reproduction, distribution or other use of this communication is strictly prohibited and could, in certain circumstances, be a criminal offense. If you have received this e-mail in error, please notify the sender by reply and delete this message without copying or disclosing it. -- 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
Re: Synchronize Time Between Mainframe and Servers?
Well, I think we've found a simple way out of this. The PC our console automation runs on is already sync'd with the network, which gets its time from an external source. For the time being, we are going to set up a script to capture this time and issue console SET CLOCK commands. May be off by a few seconds, but good enough for our humble purposes at this time. In the future, we will consider STP and sync to the same source as the network, which might eventually evolve to option 1, below (those of you who have fought in the Architecture Wars know you have to choose your battles carefully). Timothy Sipples of IBM wrote: At present you have two choices for keeping mainframe time in sync with other servers' time clocks: 1. Make the mainframe boss. This is the best choice for a lot of reasons. Set up the mainframe to get its external time reference, then use the built-in time server so that other servers can get time sync over the network, perhaps in a hierarchy. 2. Have both the mainframe and your organization's master time server (if they are unfortunately different entitites) both sync to the same external time reference (such as NIST dial-up). This might not offer as much accuracy, but it might be sufficient for certain purposes. Again, thanks to all who contributed. -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Thu, 24 May 2007 14:51:59 +0200, R.S. wrote: That means buy 9037 or STP (it's not free) and connect it to every CPC you have. I don't know STP prices, but 9037 costed approx. 50k$. You need at least two for redundancy. You don't NEED two timers unless you have a requirement for fault tolerance within a sysplex environment. If I remember correctly, you need at least one in a sysplex environment. What happens if it dies is better left to someone who knows. We aren't sysplex and can deal with the headaches if our single timer dies, which basically means you fall back to the old method of relying on the internal clocks only with no synchronization to the outside world. After reading Timothy's last comment though, it reads like the new steering mechanism will go through the HMC. And Radoslaw states that the software to do it isn't free. It's time to make an appointment with the manuals again. Has anyone actually done this and can state for certain what is involved ??? -- 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
Re: Synchronize Time Between Mainframe and Servers?
Radoslaw If it's still possible to find a couple of platforms which retain the superb AnyNet Sockets over SNA function, that problem is solved. Maybe it's still available in the Communications Server products other than z/OS, perhaps a couple of 2217s. Unfortunately the AnyNet Sockets over SNA baby got thrown out with the bathwater of prejudice against SNA. Chris Mason - Original Message - From: R.S. [EMAIL PROTECTED] Newsgroups: bit.listserv.ibm-main To: IBM-MAIN@BAMA.UA.EDU Sent: Thursday, May 24, 2007 2:51 PM Subject: Re: Synchronize Time Between Mainframe and Servers? ... BTW: SNA connection does not allow to use mainframe as time source. I heard about installations, where IP network is strictly isolated from mainframe (although it has its own IP network). Of course this is not a problem with mainframe itself. I can be a problem of any machine isolated by someone's decision. -- Radoslaw Skorupka Lodz, Poland -- 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
Re: Synchronize Time Between Mainframe and Servers?
Getting money for ETR or STP! Craig -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Shane Ginnane Sent: Wednesday, May 23, 2007 7:34 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? I can't believe this thread is still alive. Everyone (presumably) has an in-house time server for their network - possibly excluding the mainframe. Go connect your mainframe (ETR or STP) to an atomic source and let it stabilize (if you don't have either, you'd better start looking at the latter). Then either use *it* as your in-house NTP time-server, or put it as the first server on the list for the box that currently dials out for the time. What's so hard ???. Shane ... -- 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 -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Thu, 24 May 2007 08:36:47 -0500, Miller, Pat [EMAIL PROTECTED] wrote: ..., we are going to set up a script to capture this time and issue console SET CLOCK commands. May be off by a few seconds, but good enough for our humble purposes at this time. This will get the local time correct, probably better than a few seconds, perhaps even a fraction of a second. But it leaves GMT, the Gold Standard for critical timestamps drifting aimlessly. I've wondered about the possibility of a process which would make small adjustments repeatedly to CVTLSO. This should adjust both local and GMT, leaving out only jobs that issue STCK directly but ignore CVTLSO. In the future, we will consider STP and sync to the same source as the network, which might eventually evolve to option 1, below (those of you who have fought in the Architecture Wars know you have to choose your battles carefully). The first UNIX system on my desktop was a Sparc (I forget the model) running SunOS 4.3. It had a privileged command (and I had the privlege), date -a which was documented to tweak the clock frequency by a small amount until the specified time adjustment was complete, then restore it to standard rate. The man page said, a fraction of one percent. My experiment showed it was more like 25%. Whatever. It worked, and I don't believe the clock was ever observed to run backward. The description of the hardware for STP mentions similar facilities. There's a clock offset register, which is added to the value returned by STCK, and a rate offset register which is periodically added to the clock offset register. The maximum slew is comparable to that of the Sysplex Timer, a few seconds a day, not enough for some purposes. While a 25% overdrive or underdrive of the TOD clock would be considered harmful in many cases, surely it would be a boon to the Eastern Hemisphere site which is desperately seeking a way to convert from TOD=Local to TOD=GMT without shutting down for several hours. -- gil -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Thu, 24 May 2007 12:30:26 -0400, Kittendorf, Craig wrote: Getting money for ETR or STP! Is the hardware to support STP likely to be standard on future models of z/Series? If so, there's an opportunity for ISV STP lookalikes. -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Shane Ginnane What's so hard ???. -- gil -- 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
Re: Synchronize Time Between Mainframe and Servers?
-Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Paul Gilmartin Sent: Thursday, May 24, 2007 12:15 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? On Thu, 24 May 2007 12:30:26 -0400, Kittendorf, Craig wrote: Getting money for ETR or STP! Is the hardware to support STP likely to be standard on future models of z/Series? If so, there's an opportunity for ISV STP lookalikes. The STP function is microcode and perhaps some physical stuff, and I'd bet that the feature code for STP only does some microcode upgrade to enable already existing hardware. Just like with CPACF. The CPACF hardware is there, the feature code only enables it (for free, so why isn't it just standard). Or is CPACF only microcode with no supporting hardware (I don't know!). Anyway, not likely that an ISV could do a microcode upgrade. Imagine the fun that CA would have if they could install their own microcode modifications (no offense meant to CA, it is just that their software is the only stuff that I know that actually goes in and dynamically ZAPs and front-ends parts of z/OS). -- John McKown Senior Systems Programmer HealthMarkets Keeping the Promise of Affordable Coverage Administrative Services Group Information Technology The information contained in this e-mail message may be privileged and/or confidential. It is for intended addressee(s) only. If you are not the intended recipient, you are hereby notified that any disclosure, reproduction, distribution or other use of this communication is strictly prohibited and could, in certain circumstances, be a criminal offense. If you have received this e-mail in error, please notify the sender by reply and delete this message without copying or disclosing it. -- 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
Re: Synchronize Time Between Mainframe and Servers?
I think the change is on the HMC or system element and I'm not sure that is an area ISVs can get into. -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Paul Gilmartin Sent: Thursday, May 24, 2007 1:15 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? On Thu, 24 May 2007 12:30:26 -0400, Kittendorf, Craig wrote: Getting money for ETR or STP! Is the hardware to support STP likely to be standard on future models of z/Series? If so, there's an opportunity for ISV STP lookalikes. -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Shane Ginnane What's so hard ???. -- gil -- 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 -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Thu, 2007-05-24 at 12:15 -0500, Paul Gilmartin wrote: Is the hardware to support STP likely to be standard on future models of z/Series? If so, there's an opportunity for ISV STP lookalikes. If you have stand-alone CEC(s), looks like you simply need a modem; no doubt analogue. We might all have to talk to Radoslaw about sourcing those. For (parallel) sysplex, all the plumbing is in already place - CF ports and links. Then there is merely the small matter of licensing - no idea of the cost. Shane ... -- 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
Re: Synchronize Time Between Mainframe and Servers?
According to the Redbook I mentioned, each STP-capable CEC and CF must have the STP feature code 1021 installed. This feature code is a chargable feature. Your local IBM office would be able to provide you with a price quote. The book mentions that you should have the lastest microcode MCLs installed before you install the STP feature on the CECs and CFs. However before all of that work, you should probally have the HMCs upgraded to level 2.9.1 just to be safe. This upgrade appears to be a non-chargeable feature. There is a warning that after the HMC upgrade, the 'older' mainframe servers ( some 9672s and MP2000/MP3000 ) would not be manageable by that upgraded HMC. Since the ETS call out function of the STP feature uses the HMC you should already have the modem/phone line available ( does anyone run all their HMCs without a modem/phone line, just curious ). HTH Glenn Miller -- 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
Re: Synchronize Time Between Mainframe and Servers?
Glenn wrote on 25/05/2007 08:16:16 AM: Since the ETS call out function of the STP feature uses the HMC you should already have the modem/phone line available ( does anyone run all their HMCs without a modem/phone line, just curious ). Yeah, I thought about this after I posted - the modems we had to find (with RS-232-D plug) were for the timers. Hopefully for STP, there will be sufficient smarts for the current modems to do both functions. Shane ... -- 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
Re: Synchronize Time Between Mainframe and Servers?
Okay, this is proving to be one of those simple yet elusive issues. All I want to do is synchronize the time between my mainframe and the servers it exchanges data with. The servers ALREADY synchronize amongst themselves by asking the firewall server what time it is. The firewall in turn, sets its time via some internet source. I'm happy to piggyback off this arrangement. I don't really care who decides what time it is, I just care that the data that comes from a server doesn't have a timestamp of 4:37pm when the mainframe thinks it's 4:33pm. Applications apparently have trouble processing data when it appears to come from the future, even if it's only a few minutes away. The server boys say, Fine, just ask the firewall server at this IP address what time it is however often you need to in order to stay in sync. Which I would be happy to do if only SNTPD had some mechanism for me to direct it to a specific IP address address, which it apparently doesn't. This can't be that difficult, but at this point I'm cluless. -- 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
Re: Synchronize Time Between Mainframe and Servers?
Pat, STP may be your savior here. It can sync up with the same FIREWALL server or the external source that your FIREWALL uses. http://www-03.ibm.com/systems/z/pso/stp.html?ca=stpmet=inbame=WP_Site =sysz Kevin -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Miller, Pat Sent: Wednesday, May 23, 2007 12:40 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? Okay, this is proving to be one of those simple yet elusive issues. All I want to do is synchronize the time between my mainframe and the servers it exchanges data with. The servers ALREADY synchronize amongst themselves by asking the firewall server what time it is. The firewall in turn, sets its time via some internet source. I'm happy to piggyback off this arrangement. I don't really care who decides what time it is, I just care that the data that comes from a server doesn't have a timestamp of 4:37pm when the mainframe thinks it's 4:33pm. Applications apparently have trouble processing data when it appears to come from the future, even if it's only a few minutes away. The server boys say, Fine, just ask the firewall server at this IP address what time it is however often you need to in order to stay in sync. Which I would be happy to do if only SNTPD had some mechanism for me to direct it to a specific IP address address, which it apparently doesn't. This can't be that difficult, but at this point I'm cluless. -- 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 -- 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
Re: Synchronize Time Between Mainframe and Servers?
Pat, It is simple. The mainframe CANNOT synchronise its time from any external source (except from a sysplex timer). The mainframe CAN however be a source for time. So to resolve your situation everyone else needs to get their time from the mainframe (SNTPD) and not the firewall. If do that the problem will be resolved. Alan -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Miller, Pat Sent: Wednesday, May 23, 2007 11:40 To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? Okay, this is proving to be one of those simple yet elusive issues. All I want to do is synchronize the time between my mainframe and the servers it exchanges data with. The servers ALREADY synchronize amongst themselves by asking the firewall server what time it is. The firewall in turn, sets its time via some internet source. I'm happy to piggyback off this arrangement. I don't really care who decides what time it is, I just care that the data that comes from a server doesn't have a timestamp of 4:37pm when the mainframe thinks it's 4:33pm. Applications apparently have trouble processing data when it appears to come from the future, even if it's only a few minutes away. The server boys say, Fine, just ask the firewall server at this IP address what time it is however often you need to in order to stay in sync. Which I would be happy to do if only SNTPD had some mechanism for me to direct it to a specific IP address address, which it apparently doesn't. This can't be that difficult, but at this point I'm cluless. -- 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
Re: Synchronize Time Between Mainframe and Servers?
On 23 May 2007 09:46:48 -0700, in bit.listserv.ibm-main you wrote: STP may be your savior here. It can sync up with the same FIREWALL server or the external source that your FIREWALL uses. I'm a little puzzled about this statement. I just took a look at the URL you listed, and one thing I noticed is that the page makes no reference to NTP as an external source for timing information. The page does mention Initialize the time manually or by dialing out to a time service... and Schedule periodic dial-outs to a time service. Based on what I've read about STP (which I admit is not exhaustive), it looks like the z-box might not be able to time sync with the firewall if that box only supports NTP. I get the impression that an ETR (or something like an ETR) might still be required with STP. I'm sure someone will correct me if I'm wrong! :-) Eric -- Eric Chevalier E-mail: [EMAIL PROTECTED] Web: www.tulsagrammer.com Is that call really worth your child's life? HANG UP AND DRIVE! -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Wed, 23 May 2007 12:11:53 -0500, Eric Chevalier wrote: I get the impression that an ETR (or something like an ETR) might still be required with STP. I'm sure someone will correct me if I'm wrong! :-) I too am confused by this. Although in my case, that's not a particularly impressive feat. It's time to dig through the manuals some more. But an observation: IBM has discontinued the 9037 timer line. The latest version, 9037-002 is due to be dropped from support real soon (I don't remember the date). I was told that the newer z/OS releases would support STP/NTP and I guess I inferred from that, an ETR functionality. If this is not the case, how do you steer the clocks on your z server to coordinate with the rest of the world ??? Another observation: The original timer manuals and RedBooks I read were confusing and partially inaccurate, so expecting this to have changed is probably naive. -- 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
Re: Synchronize Time Between Mainframe and Servers?
I'm going to come in from the parking lot (beyond left field) on this. What cannot be done a z/OS system is moving the TOD clock backwards in time. From what I understand of the ETR and the new STP is that they adjust the speed of the TOD clock to steer it towards the correct time. Yes, this is a gross oversimplification. However, it is perfectly permissible to change the LOCAL time via a SET CLOCK=hh.mm.ss. We do this all the time on the DST change weekends. What might be acceptable, or might not, would be to have an STC which is based on the NTP code which basically issues this command as needed to adjust the LOCAL time towards the real time. I wonder how many shops would find this acceptable? The times normally shown in things like the MVS SYSLOG and JOBS is actually the LOCAL time. The time gotten via the COBOL ACCEPT verb is the LOCAL time. The only thing that I am aware of that does get the local time which is not based this way are programs which use the LE services which are sensitive to the TZ environment variable. Just a weird thought. Now I can go home! -- John McKown Senior Systems Programmer HealthMarkets Keeping the Promise of Affordable Coverage Administrative Services Group Information Technology The information contained in this e-mail message may be privileged and/or confidential. It is for intended addressee(s) only. If you are not the intended recipient, you are hereby notified that any disclosure, reproduction, distribution or other use of this communication is strictly prohibited and could, in certain circumstances, be a criminal offense. If you have received this e-mail in error, please notify the sender by reply and delete this message without copying or disclosing it. -- 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
Re: Synchronize Time Between Mainframe and Servers?
Dave, The 9037 Model 1 was withdrawn from Marketing 02/03/1997 withdrawn from Support 12/31/2003. The 9037 Model 2 was withdrawn from Marketing 12/31/2006 however, no date has been announced for the withdraw of Support. We have been looking into STP for awhile. We have found alot of planning information in the Redbook: Server Time Protocol Planning Guide SG24-7280. Regarding the ETS ( External Time Source ) question, we have found so far: If you are running in a Mixed CTN environment ( 9037s STP ), the ETS ( External Time Source ) call out would still be done by the 9037s. If you run in a STP-only CTN environment, then the ETS call out would be done by your zSeries HMC. However, the gradual adjustment to the CST ( Coordinated Server Time ) is done by the Support Element. HTH Glenn Miller -- 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
Re: Synchronize Time Between Mainframe and Servers?
Steve, From information I have seen, the expected accuracy of mainframes TOD clock is around .86 seconds per day. The spec on the 9037 ( without ETS ) seems to be much better than that, around .17 seconds per day. It seems that a mainframe that doesn't have a 9037 might see a drift of their TOD clock of as much as 25 seconds per month, yet they are still 'within spec'. If one was to experience a mainframe TOD clock drift ( current supported machines, not 3090s, ES/9000s, etc ) of something much larger than that, I recommend you place a call with IBM Hardware support. HTH Glenn Miller -- 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
Re: Synchronize Time Between Mainframe and Servers?
I can't believe this thread is still alive. Everyone (presumably) has an in-house time server for their network - possibly excluding the mainframe. Go connect your mainframe (ETR or STP) to an atomic source and let it stabilize (if you don't have either, you'd better start looking at the latter). Then either use *it* as your in-house NTP time-server, or put it as the first server on the list for the box that currently dials out for the time. What's so hard ???. Shane ... -- 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
Re: Synchronize Time Between Mainframe and Servers?
Howard Brazee wrote: On 21 May 2007 12:33:45 -0700, [EMAIL PROTECTED] (Andy Wood) wrote: The magnetron in my microwave oven cycles at a sub-nanosecond rate. That does not mean it would be be suitable as a low drift clock source. Personal computers have had built in clocks forever - but they have been cheap and unreliable. The solution was to synchronize often. I bought a program that did this calculating the lag in my modem and phone line, but nowadays it is part of the OS. First PCs wre sold without the clock. That's why MS-DOS asks for current date and time at startup (only when no autoexec.bat is present). Of course I mean battery-operated clock, working even if computer is powered off. When the PC is up, the clock is present even in those vry old XT machines. However the clock is inaccurate by design. The time is measured in ticks. There are 64k per hour. Due to round erors the time is always drifting. -- Radoslaw Skorupka Lodz, Poland -- BRE Bank SA ul. Senatorska 18 00-950 Warszawa www.brebank.pl Sd Rejonowy dla m. st. Warszawy XII Wydzia Gospodarczy Krajowego Rejestru Sdowego, nr rejestru przedsibiorców KRS 025237 NIP: 526-021-50-88 Wedug stanu na dzie 01.01.2007 r. kapita zakadowy BRE Banku SA (w caoci opacony) wynosi 118.064.140 z. W zwizku z realizacj warunkowego podwyszenia kapitau zakadowego, na podstawie uchwa XVI WZ z dnia 21.05.2003 r., kapita zakadowy BRE Banku SA moe ulec podwyszeniu do kwoty 118.760.528 z. Akcje w podwyszonym kapitale zakadowym bd w caoci opacone. -- 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
Re: Synchronize Time Between Mainframe and Servers?
Thompson, Steve wrote: [...] The magnetron in my microwave oven cycles at a sub-nanosecond rate. That does not mean it would be be suitable as a low drift clock source. snip Would that be because it does not have any missing-beat detection micro-code (no pun intended) and would as a result post an error condition (machine check)? Perhaps there is no circuitry to force the magnetron to lock at a precise frequency (say 1.037 GHz +/- 10KHz) because the freq drift is (a) not important as long as you stay around the assigned micro-wave freq +/- 1% so that leakage is not a problem? [...] I don't know microwave owen design, but I can imagine a device working at - let's say - 3,5GHz +/- 20%. It can be fully acceptable for the device (like owen) to work at 4.2GHz as well as 2,8GHz. So, resolution is sub-nanosecond, but the accuracy is very poor. From the other hand quarts oscillators usually work at MHz frequency, but the frequency is very accurate and stable over time (months, years). -- Radoslaw Skorupka Lodz, Poland -- BRE Bank SA ul. Senatorska 18 00-950 Warszawa www.brebank.pl Sd Rejonowy dla m. st. Warszawy XII Wydzia Gospodarczy Krajowego Rejestru Sdowego, nr rejestru przedsibiorców KRS 025237 NIP: 526-021-50-88 Wedug stanu na dzie 01.01.2007 r. kapita zakadowy BRE Banku SA (w caoci opacony) wynosi 118.064.140 z. W zwizku z realizacj warunkowego podwyszenia kapitau zakadowego, na podstawie uchwa XVI WZ z dnia 21.05.2003 r., kapita zakadowy BRE Banku SA moe ulec podwyszeniu do kwoty 118.760.528 z. Akcje w podwyszonym kapitale zakadowym bd w caoci opacone. -- 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
Re: Synchronize Time Between Mainframe and Servers?
-Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of R.S. Sent: Tuesday, May 22, 2007 1:20 AM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? SNIP I don't know microwave owen design, but I can imagine a device working at - let's say - 3,5GHz +/- 20%. It can be fully acceptable for the device (like owen) to work at 4.2GHz as well as 2,8GHz. So, resolution is sub-nanosecond, but the accuracy is very poor. From the other hand quarts oscillators usually work at MHz frequency, but the frequency is very accurate and stable over time (months, years). SNIP Yes, we used to call that rock bound (Crystals) [I haven't had an FCC license at commercial or amateur level since about 1977 -- things change.] I'm told, but have no experience with PLL (Phase Locked Loop), that it is also as stable and into the GHz range. But let us get back on topic. The Mainframe's (at least from IBM's design), has a TOD that if it malfunctions, is detected and presented as a machine check. PCs, as far as I've been able to deduce, do not. As a result, the mainframe TOD has a design that prevents drift (or at least it did in the TTL days). So much so that in working on ETR designs it was questioned as to how dithering could be done without causing TOD error code giving failures. Obviously that microcode had to be changed to allow an external timer to make the changes w/o causing Clock Comparator errors (among others). Regards, Steve Thompson -- STD disclaimer: Poster's opinions are not necessarily those of poster's employer. -- -- 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
Re: Synchronize Time Between Mainframe and Servers?
-Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of R.S. Sent: Saturday, May 19, 2007 8:22 AM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? As usually in mainframe world, we have a lot of proofs that mainframe is better, we can provide citations from ESA/390 Holy Principles, we can answer that we have ntp server... But the only honest answer to the question is NO, unfortunately we are not able to synchronize with atomic clock, like other devices in our server room. Mainframe *can't do it*. We can only use wrist watch as a time source and then rely on mainframe internal clock. Of course, after the confession we can go back to our Principles, talk about our availability, hint that the rest of the world should synchronize with our CPC... -- Radoslaw Skorupka Lodz, Poland P.S. I would like to hear that I'm wrong. Preferrably with details - how to synchronize the mainframe with ntp server or atomic clock. Preferrably without 9037 Timers, which are expensive and EOM (End Of Market). snip First, I was answering the original question as the poster asked it, with the understanding that they did NOT have an ETR (v1 or v2). As late as 2003 I had a problem with a client who migrated off VSE to z/OS, and then tried to migrate off the mainframe to other platforms. The need for time sync was demonstrable (DR purposes was secondary!), and we could not seem to get management of the company to recognize the issues (they bought the idea that NTP was sufficient). The technical people handling the non-mainframe platforms did understand, and the DBAs (non-mainframe) recognized that time sync using the mainframe would be much more reliable than what they were trying to do (even with that installation using an operator's view of a clock pending the strike of the enter key). So sometimes what you have is a purely political issue. Second, you most certainly can use an external time source to set the mainframe (either via Unix System Services TCP/IP, or through the HMC -- neither of which I have personal experience doing). The question is, since the time source you will attempt to sync to (absent an ETR connected to a GPS/UTC time source) is NOT sufficiently accurate to avoid ambiguities in TOD processing -- why would you desire to use NTP to set the mainframe, as opposed to letting the mainframe be your NTP source? This was my reason for pointing out, from the PoOP, that the TOD cycles bit 51 at the micro-second rate, giving a very small drift (baring a bad power source as someone else pointed out -- but that kind of issue should cause the system to crash because of, shall we call them timer-checks?) compared to Intel Architecture based clocks, which seem to suffer significant drift problems (I have personal experience in this case using Tardis to keep a LAN synchronized and looking at the drift report on the NT server was SCARY - 2-10 seconds in a day, and other LANs with other software/hardware with similar issues). The current thing coming from IBM (which I greatly appreciate and can't wait until we take delivery on the new hardware to use it) is STP. But not having ANY experience with it yet, I can't truly speak to it. Regards, Steve Thompson -- STD disclaimer: poster's posting may not reflect poster's employer's opinions and should not necessarily be construed as such. Deputy Dawg did not approve of this message. YMMV -- -- 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
Re: Synchronize Time Between Mainframe and Servers?
Thanks to all who have commented. It's been an education. SNTPD started task. Seems simple enough. Howsomever, neither the IP Config guide nor Reference tells you much more than the name of the procedure, where to find the sample, and what it does. I.e., neither tells you how it works or any setup you have to provide in order for it to key off an ETR. Surely I have to at least point something to the ETR and/or provide some authentication? -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Field, Alan C. Sent: Friday, May 18, 2007 4:06 PM To: IBM-MAIN@BAMA.UA.EDU Subject:Re: Synchronize Time Between Mainframe and Servers? Pat, 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. Have your enterprise query the SNTP on the mainframe and set their clocks off that for a consistent timestamp across your organization. Developers are asking me if we can do this to prevent anomalies between timestamps across platforms. In case it's an issue, no, I don't have a sysplex or timer. -- 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 -- 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
Re: Synchronize Time Between Mainframe and Servers?
Pat, It is that simple. My guess is that SNTPD task queries the time in the processor clock (think TIME or STCK macro) and I presume formats the result in an appropriate NTP format and makes it available on the appropriate port (123?) to all requesters. If you have an ETR AND you set and correct it using an external source (like NIST) then you will be able to provide accurate time to your organization. If you set the mainframe clock with the operator's wristwatch then you might not have accurate time but you will have consistent timestamps across all platforms which is what you are requesting i.e. no more transactions completing on one platform before they have begun on another. Alan -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Miller, Pat Thanks to all who have commented. It's been an education. SNTPD started task. Seems simple enough. Howsomever, neither the IP Config guide nor Reference tells you much more than the name of the procedure, where to find the sample, and what it does. I.e., neither tells you how it works or any setup you have to provide in order for it to key off an ETR. Surely I have to at least point something to the ETR and/or provide some authentication? Pat, 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. Have your enterprise query the SNTP on the mainframe and set their clocks off that for a consistent timestamp across your organization. Developers are asking me if we can do this to prevent anomalies between timestamps across platforms. In case it's an issue, no, I don't have a sysplex or timer. -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Fri, 18 May 2007 18:19:20 -0400, Thompson, Steve wrote: 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. Our old z800 would drift a few seconds every month. I haven't had the opportunity to observe the accuracy of the z9 clocks. But the 9037-002 is going obsolete soon, so I'm going to have to address this again. -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Sat, 19 May 2007 17:02:52 +0200, R.S. wrote: 1. I don't know whether STP solution can use 'atomic clock' as a time reference. AFAIK, it was asked on the list, but noone answered. 2. (**) Two sysplex timers are required, because of redundancy. AFAIK if you loose ETR (sysplex timer or timer connection), then all your systems will end with wait state. We only have one timer. We only use it to keep the clocks relatively close to reality and to keep them consistent across the enterprise. Also, if you lose the timer, you can recover and run without it. It takes a few seconds for error recovery and to present the option on the console. By this time, most of our communication sessions time out, so we've got some work to do to get everything reconnected. But batch JOBs continue to run and everything else runs just fine too once you restart all your VTAM applications and such. -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Mon, 21 May 2007 09:55:13 -0400, Thompson, Steve [EMAIL PROTECTED] wrote: . . . Second, you most certainly can use an external time source to set the mainframe (either via Unix System Services TCP/IP, or through the HMC -- neither of which I have personal experience doing). The question is, since the time source you will attempt to sync to (absent an ETR connected to a GPS/UTC time source) is NOT sufficiently accurate to avoid ambiguities in TOD processing -- why would you desire to use NTP to set the mainframe, as opposed to letting the mainframe be your NTP source? This was my reason for pointing out, from the PoOP, that the TOD cycles bit 51 at the micro-second rate, giving a very small drift . . . The magnetron in my microwave oven cycles at a sub-nanosecond rate. That does not mean it would be be suitable as a low drift clock source. -- 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
Re: Synchronize Time Between Mainframe and Servers?
-Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Andy Wood Sent: Monday, May 21, 2007 2:29 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? On Mon, 21 May 2007 09:55:13 -0400, Thompson, Steve [EMAIL PROTECTED] wrote: . . . Second, you most certainly can use an external time source to set the mainframe (either via Unix System Services TCP/IP, or through the HMC -- neither of which I have personal experience doing). The question is, since the time source you will attempt to sync to (absent an ETR connected to a GPS/UTC time source) is NOT sufficiently accurate to avoid ambiguities in TOD processing -- why would you desire to use NTP to set the mainframe, as opposed to letting the mainframe be your NTP source? This was my reason for pointing out, from the PoOP, that the TOD cycles bit 51 at the micro-second rate, giving a very small drift . . . The magnetron in my microwave oven cycles at a sub-nanosecond rate. That does not mean it would be be suitable as a low drift clock source. snip Would that be because it does not have any missing-beat detection micro-code (no pun intended) and would as a result post an error condition (machine check)? Perhaps there is no circuitry to force the magnetron to lock at a precise frequency (say 1.037 GHz +/- 10KHz) because the freq drift is (a) not important as long as you stay around the assigned micro-wave freq +/- 1% so that leakage is not a problem? (b) H2O doesn't necessarily care about the precise frequency, close enough will do? (c) the FCC (or equivalent) doesn't care as long as there is insufficient leakage to interfere with other SHF devices? Sorry, but I've retired from working on radar and micro-wave stuff and haven't touched it since solid state took over. Regards, Steve Thompson -- Std. Disclaimer about employer, opinions, etc. -- -- 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
Re: Synchronize Time Between Mainframe and Servers?
On 21 May 2007 12:33:45 -0700, [EMAIL PROTECTED] (Andy Wood) wrote: The magnetron in my microwave oven cycles at a sub-nanosecond rate. That does not mean it would be be suitable as a low drift clock source. Personal computers have had built in clocks forever - but they have been cheap and unreliable. The solution was to synchronize often. I bought a program that did this calculating the lag in my modem and phone line, but nowadays it is part of the OS. I wonder how Vax's used to do this with their early acceptance of world wide systems being the norm. -- 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
Re: Synchronize Time Between Mainframe and Servers?
In !!AAAYAN4tmqr/aItLkw6fT3JvZv/[EMAIL PROTECTED], on 05/18/2007 at 09:55 PM, Doc Farmer [EMAIL PROTECTED] said: But then again, I'm a bit of a horophile (no sniggering, it doesn't mean THAT) I don't see what else it could mean; I can't think of any other Greek or Latin word with the same root. But why would I find a reference to time to be salacious in the first place? -- Shmuel (Seymour J.) Metz, SysProg and JOAT ISO position; see http://patriot.net/~shmuel/resume/brief.html We don't care. We don't have to care, we're Congress. (S877: The Shut up and Eat Your spam act of 2003) -- 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
Re: Synchronize Time Between Mainframe and Servers?
In [EMAIL PROTECTED], on 05/19/2007 at 03:22 PM, R.S. [EMAIL PROTECTED] said: As usually in mainframe world, we have a lot of proofs that mainframe is better, we can provide citations from ESA/390 Holy Principles, we can answer that we have ntp server... But the only honest answer to the question is NO, unfortunately we are not able to synchronize with atomic clock, like other devices in our server room. Mainframe *can't do it*. We can only use wrist watch as a time source and then rely on mainframe internal clock. No, that is not an honest answer. An honest answer would be that you are currently running without the required hardware and software but that it has been possible for lo these many releases when correctly configured. P.S. I would like to hear that I'm wrong. Okay, you're wrong. Preferrably without 9037 Timers, And you knew that you were wrong. No doubt you have an objection to STP as well, but the fact that you can't or w'ont use it doesn't mean that it's not available. -- Shmuel (Seymour J.) Metz, SysProg and JOAT ISO position; see http://patriot.net/~shmuel/resume/brief.html We don't care. We don't have to care, we're Congress. (S877: The Shut up and Eat Your spam act of 2003) -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Fri, 18 May 2007 15:37:44 -0500, Miller, Pat [EMAIL PROTECTED] wrote: Developers are asking me if we can do this to prevent anomalies between timestamps across platforms. In case it's an issue, no, I don't have a sysplex or timer. Well we do it in my shop ( i started with Ken Clapp code first) Now we use the SNTP daemon provided by TCPIP , and all hardware machines are using it .(we run it on a // sysplex dual site) For example the AD domain server is synchronising once per hour , and the 2500 clients are synchronising to the domain server once or twice a day . All 250 servers ( W2k/Linux and AIX) the Cisco , Nortel routers are also synchronised once or twice a day to the SNTP server . Our 400 printers were also synchronised , but we just changed the brand , and i do not know it it is still the case . The SNTP adress is a Dynamic VIPA vith a Vipabackup on the second CPU of the sysplex ( in order to keep answering to AD request during an IPL for example) This is quite good for debugging as logs are coherent . In clear the time is perhaps wrong , but it is the same for everyone ! Bruno ( on private mail at home) Bruno(dot)sugliani(at)groupemornay(dot)asso(dot)fr -- 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
Re: Synchronize Time Between Mainframe and Servers?
Any DB2 recovery I performed was either to an IMAGCOPY or RBA value. I may be rusty, but there was not 'to a point of time' recovery. You can capture an RBA value from a point in time, but you use the RBA value. IMS keeps RBA values of logs, too. In a SYSPLEX, you either run on the same box so the time is always the same or you have to have a Sysplex Timer (going away so start replacing them) or the new System Time Protocol (STP) feature in the newer boxes. Eric, I think (although do not know for sure) that it become important for logs like DB2 for recovery purposes. I think that was the talk on here a while ago. Someone please pipe up if I am wrong. Ed -- 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
Re: Synchronize Time Between Mainframe and Servers?
If you're off a second or three, you're right - no biggie. However, if you've got a multiple-mainframe environment and you're running time-sensitive transactions, it's best to keep their time in sync as much as possible. 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. Perception. Corporate LAN runs off to a timesource every so often - all the users ever see is a consistent (correct) time value. Mainframe (even with ETR) wanders around always off-time - unless it also synchs to a (different) timesource. Given the questions we see here on the list, I wonder if the majority of ETRs aren't synched to an atomic source at all, but set locally. Is it not time to start plugging those ETRs into a phone jack and using the technology that will correct the time? Go to the SHARE Conference or start researching Server Time Protocol (STP) in the new boxes. NTP is not accurate enough for the amount of work a mainframe can handle. The weenie boxes may be okay processing with a few seconds difference, but Parallel Sysplex can not. Sysplex timers can dial out to a timesource so the operator's wrist watch was only a starting point. And if Corporate allows time to be set that way, shame on them. Let your wan/lan get the time off the mainframe and you will be using a consistent time. Replacing Sysplex Timers: Server Time Protocol (STP) Overview and Planning Part 1 of 2 - 08/15/2007 08:00 Wednesday In this session, part one of two, the speakers will discuss an overview of the Server Time Protocol (STP), the new paradigm for synchronizing clocks between System z servers in a timing network. -- 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
Re: Synchronize Time Between Mainframe and Servers?
Any DB2 recovery I performed was either to an IMAGCOPY or RBA value. I may be rusty, but there was not 'to a point of time' recovery. You can capture an RBA value from a point in time, but you use the RBA value. IMS keeps RBA values of logs, too. Unfortunately, you are rusty... When you go to an IMSPLEX, or a DB2PLEX, the logs are running for each member. So, the RBA for each instance has no correlation with another. You have to use timestamps to recover, back-out, roll forward for any failure of any shared data. - Too busy driving to stop for gas! -- 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
Re: Synchronize Time Between Mainframe and Servers?
As usually in mainframe world, we have a lot of proofs that mainframe is better, we can provide citations from ESA/390 Holy Principles, we can answer that we have ntp server... But the only honest answer to the question is NO, unfortunately we are not able to synchronize with atomic clock, like other devices in our server room. Mainframe *can't do it*. We can only use wrist watch as a time source and then rely on mainframe internal clock. Of course, after the confession we can go back to our Principles, talk about our availability, hint that the rest of the world should synchronize with our CPC... -- Radoslaw Skorupka Lodz, Poland P.S. I would like to hear that I'm wrong. Preferrably with details - how to synchronize the mainframe with ntp server or atomic clock. Preferrably without 9037 Timers, which are expensive and EOM (End Of Market). -- BRE Bank SA ul. Senatorska 18 00-950 Warszawa www.brebank.pl Sd Rejonowy dla m. st. Warszawy XII Wydzia Gospodarczy Krajowego Rejestru Sdowego, nr rejestru przedsibiorców KRS 025237 NIP: 526-021-50-88 Wedug stanu na dzie 01.01.2007 r. kapita zakadowy BRE Banku SA (w caoci opacony) wynosi 118.064.140 z. W zwizku z realizacj warunkowego podwyszenia kapitau zakadowego, na podstawie uchwa XVI WZ z dnia 21.05.2003 r., kapita zakadowy BRE Banku SA moe ulec podwyszeniu do kwoty 118.760.528 z. Akcje w podwyszonym kapitale zakadowym bd w caoci opacone. -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Sat, 19 May 2007 15:22:14 +0200, R.S. [EMAIL PROTECTED] wrote: But the only honest answer to the question is NO, unfortunately we are not able to synchronize with atomic clock, like other devices in our server room. Mainframe *can't do it*. We can only use wrist watch as a time source and then rely on mainframe internal clock. Of course, after the confession we can go back to our Principles, talk about our availability, hint that the rest of the world should synchronize with our CPC... What do you mean ? Why don't you synchronise the syslex timer with atomic clock or ( easier and cheaper,) radio longwave transmission clock ? ( this one does not even require an external antenna because it crosses the walls of you computer room and is used for railways , tunels etc ... ) Or did i misunderstood your comments ? Bruno Bruno(dot)sugliani(at)groupemornay(dot)asso(dot)fr -- 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
Re: Synchronize Time Between Mainframe and Servers?
Bruno Sugliani wrote: On Sat, 19 May 2007 15:22:14 +0200, R.S. [EMAIL PROTECTED] wrote: But the only honest answer to the question is NO, unfortunately we are not able to synchronize with atomic clock, like other devices in our server room. Mainframe *can't do it*. We can only use wrist watch as a time source and then rely on mainframe internal clock. Of course, after the confession we can go back to our Principles, talk about our availability, hint that the rest of the world should synchronize with our CPC... What do you mean ? Why don't you synchronise the syslex timer with atomic clock or ( easier and cheaper,) radio longwave transmission clock ? ( this one does not even require an external antenna because it crosses the walls of you computer room and is used for railways , tunels etc ... ) Or did i misunderstood your comments ? 1. I wrote 'atomic clock', but I meant any 'external time reference' including atomic clock, and longwave transmission. In fact I even don't know what's available in our part of Europe. For sure, there is something, because we have devices, 'time receivers'. However it has little to do with mainframe, since it cannot use any of the time sources. 2. AFAIK such synchronization requires some special setup, sysplex timer will not start using it itself. How can I connect sysplex timer to my ntp server over regular IP/ethernet connection ? 3. Last, but definitely not least: I DON'T HAVE sysplex timer. Sysplex timer (I wrote about it in post scriptum), are no longer available from IBM, so there is no supported way to connect to time reference. Also the price is killing factor: two (**) sysplex timers costed approx. 100k$ !!! It is too much if you simply want synchronize your machine with other machines, isn't it ? 1. I don't know whether STP solution can use 'atomic clock' as a time reference. AFAIK, it was asked on the list, but noone answered. 2. (**) Two sysplex timers are required, because of redundancy. AFAIK if you loose ETR (sysplex timer or timer connection), then all your systems will end with wait state. -- Radoslaw Skorupka Lodz, Poland -- BRE Bank SA ul. Senatorska 18 00-950 Warszawa www.brebank.pl Sd Rejonowy dla m. st. Warszawy XII Wydzia Gospodarczy Krajowego Rejestru Sdowego, nr rejestru przedsibiorców KRS 025237 NIP: 526-021-50-88 Wedug stanu na dzie 01.01.2007 r. kapita zakadowy BRE Banku SA (w caoci opacony) wynosi 118.064.140 z. W zwizku z realizacj warunkowego podwyszenia kapitau zakadowego, na podstawie uchwa XVI WZ z dnia 21.05.2003 r., kapita zakadowy BRE Banku SA moe ulec podwyszeniu do kwoty 118.760.528 z. Akcje w podwyszonym kapitale zakadowym bd w caoci opacone. -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Sat, 19 May 2007 02:11:01 -0500, Bruno Sugliani wrote: Well we do it in my shop ( i started with Ken Clapp code first) Now we use the SNTP daemon provided by TCPIP , and all hardware machines are using it .(we run it on a // sysplex dual site) ... In clear the time is perhaps wrong , but it is the same for everyone ! What are the z/Series using as a time reference? Is it possible for one z/Series to use the SNTP daemon on another z/Series system as a reference? -- gil -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Sat, 19 May 2007 17:02:52 +0200, R.S. wrote: 2. AFAIK such synchronization requires some special setup, sysplex timer will not start using it itself. How can I connect sysplex timer to my ntp server over regular IP/ethernet connection ? I have heard of a scheme where an NTP-synchronized PC is connected via a null modem to the serial port on the sysplex timer (the PS/2 in the sysplex timer can't do this by itself). That PC then spoofs the protocol of dialup to NIST, which is aceptable to the sysplex timer. Rube Goldberg. I can imagine hearing the PFCSKs snickering. When they learn the price of the obsolete PC which is the sysplex timer, the snickers become guffaws. -- gil -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Sat, 19 May 2007 10:12:04 -0500, Paul Gilmartin [EMAIL PROTECTED] wrote: What are the z/Series using as a time reference? Is it possible for one z/Series to use the SNTP daemon on another z/Series system as a reference? The z/Series are using the 9037 as an external source and one of them become the server for the distributed servers or routers ( they are clients calling on UDP port 123) The 9037 can get its time from longwave or atomic or whatever (via the V24) . But i do not know of an SNTP client running on z/Series , so i do not think it is possible . ( we could make a nice loop however :-)) ) I am no expert ( i was just fed up having different timestamps on all multiple tiers architecture , and now it is solved ) . Bruno -- 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
Re: Synchronize Time Between Mainframe and Servers?
Paul Gilmartin wrote: What are the z/Series using as a time reference? Is it possible for one z/Series to use the SNTP daemon on another z/Series system as a reference? Options (I think they were mostly OEM) have existed for some time to connect a Sysplex Timer with an ETR such as a receiver for a Radio transmitter like the German DCF77 or a GPS receiver. To my knowledge, there hasn't been much customer interest in these devices. Using SNTP to get the time might be easier but it is, I think, understandable that installations are careful about connecting a Sysplex Timer to the Internet. The problem of adjusting the time base of a sysplex is quite complex. Increments have to be chosen very carefully and simply setting the time back is out of the question. -- Ulrich Boche SVA GmbH, Germany IBM Premier Business Partner -- 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
Re: Synchronize Time Between Mainframe and Servers?
Pat, 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. Have your enterprise query the SNTP on the mainframe and set their clocks off that for a consistent timestamp across your organization. Developers are asking me if we can do this to prevent anomalies between timestamps across platforms. In case it's an issue, no, I don't have a sysplex or timer. -- 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
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? Have your enterprise query the SNTP on the mainframe and set their clocks off that for a consistent timestamp across your organization. This may be politically infeasible. It's pretty hard to rebut the argument, Why shouldn't the mainframe adjust itself to the correct time elsewhere in of the enterprise rather than demanding that the remainder of the enterprise adjust to the mainframe's incorrect time? with a blunt It can't. Will STP provide a solution? Developers are asking me if we can do this to prevent anomalies between timestamps across platforms. In case it's an issue, no, I don't have a sysplex or timer. -- gil -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Fri, 2007-05-18 at 16:28 -0500, Paul Gilmartin wrote: Will STP provide a solution? I would hope STP will make the mainframe (server, did some-one say server ???) a little less intransigent in this regard. It'll dial out and correct itself - just make it the corporate timesource. Would be an idea to pick a long-running LPAR for SNTP though ... And a CEC that's on UTC ... And get the LAN folks convinced it worthwhile changing ... Piece of cake. Shane ... -- 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
Re: Synchronize Time Between Mainframe and Servers?
-Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Paul Gilmartin Sent: Friday, May 18, 2007 4:29 PM To: IBM-MAIN@BAMA.UA.EDU 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
Re: Synchronize Time Between Mainframe and Servers?
On Fri, 2007-05-18 at 18:19 -0400, Thompson, Steve wrote: 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. Perception. Corporate LAN runs off to a timesource every so often - all the users ever see is a consistent (correct) time value. Mainframe (even with ETR) wanders around always off-time - unless it also synchs to a (different) timesource. Given the questions we see here on the list, I wonder if the majority of ETRs aren't synched to an atomic source at all, but set locally. It's about time IBM allowed the clock correction to be driven by an accepted source. Maybe the next step will be to just be a (local) client like everyone else. It's just a server after all ... Shane ... -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Sat, 19 May 2007 08:42:28 +1000, Shane [EMAIL PROTECTED] wrote: On Fri, 2007-05-18 at 18:19 -0400, Thompson, Steve wrote: 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. Perception. Corporate LAN runs off to a timesource every so often - all the users ever see is a consistent (correct) time value. Mainframe (even with ETR) wanders around always off-time - unless it also synchs to a (different) timesource. Given the questions we see here on the list, I wonder if the majority of ETRs aren't synched to an atomic source at all, but set locally. It's about time IBM allowed the clock correction to be driven by an accepted source. Maybe the next step will be to just be a (local) client like everyone else. It's just a server after all ... Shane ... -- 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 -- 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
Re: Synchronize Time Between Mainframe and Servers?
On Sat, 19 May 2007 08:42:28 +1000, Shane wrote: On Fri, 2007-05-18 at 18:19 -0400, Thompson, Steve wrote: Corporate LAN runs off to a timesource every so often - all the users ever see is a consistent (correct) time value. Mainframe (even with ETR) wanders around always off-time - unless it also synchs to a (different) timesource. Given the questions we see here on the list, I wonder if the majority of ETRs aren't synched to an atomic source at all, but set locally. Everything runs off to a timesource every so often. Even the atomic source is a weighted average of several atomic clocks widely separated geographically. I think that's what the coordinated in UTC means. It's dismaying that for mainframes the timesource is too often the operator's wristwatch. It's about time IBM allowed the clock correction to be driven by an accepted source. I believe that's what the sysplex timer does. In my experience, it does it extremely well. I don't know whether dialup or shortwave to WWV is counted as an accepted source. Alas, the OP states he hasn't a Sysplex Timer. ... Maybe the next step will be to just be a (local) client like everyone else. It's just a server after all ... But what happens to revenues for the Sysplex Timer division? Another perception is that time from TCP/IP is free. The source code is available at no charge. Most non-mainframe systems are sold with it preloaded. I perceive the administrative overhead as less than for a sysplex timer. As long as the mainframes remain egregious in not using TCP/IP as a timesource, there's a selling point for the competition. -- gil -- 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
Re: Synchronize Time Between Mainframe and Servers?
Unless you have some extremely time sensitive applications, if your running a sysplex and all the lpars are using the same time, whats the big deal. If you're off by less than a second or three, I doubt if most things would really matter. I'm sure there are a lot of things that would matter, but I doubt if the last 2 jobs I had - one in manufacturing and one in catalog sales, would really matter. Eric Bielefeld Sr. z/OS Systems Programmer Milwaukee, Wisconsin 414-475-7434 - Original Message - From: Shane [EMAIL PROTECTED] On Fri, 2007-05-18 at 18:19 -0400, Thompson, Steve wrote: 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. Perception. Corporate LAN runs off to a timesource every so often - all the users ever see is a consistent (correct) time value. Mainframe (even with ETR) wanders around always off-time - unless it also synchs to a (different) timesource. Given the questions we see here on the list, I wonder if the majority of ETRs aren't synched to an atomic source at all, but set locally. It's about time IBM allowed the clock correction to be driven by an accepted source. Maybe the next step will be to just be a (local) client like everyone else. It's just a server after all ... Shane ... -- 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
Re: Synchronize Time Between Mainframe and Servers?
If you're off a second or three, you're right - no biggie. However, if you've got a multiple-mainframe environment and you're running time-sensitive transactions, it's best to keep their time in sync as much as possible. If you sync to WAVY, or via Internet, or via GPS, even once a day, you're pretty much set anyway. If your mainframe clock is losing more than 2 seconds a day, you're not scrubbing your incoming power well enough (IMNSHO). Generally, I prefer to keep systems in sync as much as possible if the business needs to accurately track the transaction timings from the terminal through the network up to the mainframe and back again. It also comes in handy to have accurate time stamps for SWIFT (and even some ACH) transactions. But then again, I'm a bit of a horophile (no sniggering, it doesn't mean THAT) so I've got a variety of atomic clocks and watches all over the place. -Original Message- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Eric Bielefeld Sent: Friday, May 18, 2007 20:29 To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Synchronize Time Between Mainframe and Servers? Unless you have some extremely time sensitive applications, if your running a sysplex and all the lpars are using the same time, whats the big deal. If you're off by less than a second or three, I doubt if most things would really matter. I'm sure there are a lot of things that would matter, but I doubt if the last 2 jobs I had - one in manufacturing and one in catalog sales, would really matter. Eric Bielefeld Sr. z/OS Systems Programmer Milwaukee, Wisconsin 414-475-7434 - Original Message - From: Shane [EMAIL PROTECTED] On Fri, 2007-05-18 at 18:19 -0400, Thompson, Steve wrote: 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. Perception. Corporate LAN runs off to a timesource every so often - all the users ever see is a consistent (correct) time value. Mainframe (even with ETR) wanders around always off-time - unless it also synchs to a (different) timesource. Given the questions we see here on the list, I wonder if the majority of ETRs aren't synched to an atomic source at all, but set locally. It's about time IBM allowed the clock correction to be driven by an accepted source. Maybe the next step will be to just be a (local) client like everyone else. It's just a server after all ... Shane ... -- 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 -- 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
Re: Synchronize Time Between Mainframe and Servers?
On May 18, 2007, at 7:29 PM, Eric Bielefeld wrote: Unless you have some extremely time sensitive applications, if your running a sysplex and all the lpars are using the same time, whats the big deal. If you're off by less than a second or three, I doubt if most things would really matter. I'm sure there are a lot of things that would matter, but I doubt if the last 2 jobs I had - one in manufacturing and one in catalog sales, would really matter. Eric Bielefeld Sr. z/OS Systems Programmer Milwaukee, Wisconsin 414-475-7434 Eric, I think (although do not know for sure) that it become important for logs like DB2 for recovery purposes. I think that was the talk on here a while ago. Someone please pipe up if I am wrong. Ed -- 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