I don't think anyone means that the DS6800 has to be down for the two hours.
In my case, I leave all the systems up, but don't run anything critical (that may have backout or restore processes to consider), and let IBM do their thing. When the first controller is taken offline, we get the missing interrupts, and I think automatic retries are suppose to happen, in which case the other controller will be used. Then the microcode is loaded and the controller is IML'ed. About a 75 minute process, depending on many factors. The controller is brought online. Then the second controller is brought offline, and we get missing interrupts. The microcode is loaded and that controller is IML'ed and put back online. Even if you had a total power failure, if it messes anything up, it is only 1 controller that would be affected. The other controller will be handling all your I/Os. Not a problem unless you are doing so many I/Os that a single controller would be the bottleneck (along with it's half of the cache). If things are really critical, you might try: Bringing your systems down. Vary the controller offline. Bring your systems up. Load the microcode. IML the controller Bring the controller online. Next weekend: Bringing your systems down. Vary the other controller offline. Bring your systems up. Load the microcode. IML the controller Bring the controller online. No need for a system to be down the entire time. But if you don't know, I would leave the systems up, with as little production updating going on, and try it. On a supported OS, it should work just fine. Tom Duerbusch THD Consulting >>> Edward M Martin <emar...@aultman.com> 1/23/2009 3:08 PM >>> Hello Crispin Hugo, Mace (Larry Macioce), and Tom Duerbusch, Thank you for the info. I am on both z/VM 5.3 and z/VSE 4.1.2. I tried to wire the two racks to all have a path if we lost Full connection. It is the two hours that worry me. I get Tuesday nights here at the Hospital. And as long it works. So what is your backup plan? What happens if you get a power failure in the middle of the update? Yea, I know UPS but I have to plan for worst case. Say in your best Germany accent 'WE WILL NOT FAIL!' (sorry Martin) Crispin, you have indicated the 3 updates. I see the SEA.JAR that I believe will update the microcode in the DS6800, I see the DS6800STorageManager_WIN_5.2.200.1678.zip (for the Windows version of Storage Manager), and DSCLI_5.2.2.411.tar.gz and DSCLI_5.2.2.411.zip which I thought was the Data Storage Command Line interface. Are these the three that you updates you are talking about? Because it would be ok to get the SEA.JAR done and the do the Windows Storage Manager as the system runs. Ed Martin Aultman Health Foundation 330-588-4723 ext 40441 From: The IBM z/VM Operating System [mailto:ib...@listserv.uark.edu] On Behalf Of Crispin Hugo Sent: Friday, January 23, 2009 3:32 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: DS6800 concurrent microcode updates Our experience with VM and Z/os is that they may not survive. It appears to me to that it drops the second connection too soon after the first one is put back online. We always close system down. With our system it takes about 2 hours to do all three software updates Crispin Hugo Systems Programmer, Macro 4 <http://www.macro4.com/> ____________________
