When we were setting up our 1st Tech LPAR, TCPIP was pretty new to our mainframe. And in my knowledge. I set-up the parms for the new stack, mostly cloning and some experimentation. When looking at ROUTER/NONROUTER I got it wrong and specified ROUTER. We brought the LPAR up, all looked well and it was the end of the day. A little later, at the grocery store I got a call, why as the entire campus's traffic trying to pass through my capped (quite low) new test LPAR? And basically, all traffic as at a standstill. Qur main TCPIP backbone of the time saw my new stack with ROUTER, believed it and saw a shorter path. >From the grocery store, I had the operator issue a QUIESCE on the new LPAR >console, and the backbone saw it become nonresponsive and things got better. Twas an interesting post-mortem meeting the next day. I learned more, and the backbone was updated to not accept just any ROUTER claimant. > -----Original Message----- > From: IBM Mainframe Discussion List <[email protected]> On > Behalf Of Radoslaw Skorupka > Sent: Friday, March 7, 2025 8:57 AM > To: [email protected] > Subject: Rescue system (was stupid outages) > > Some of us wrote about so called rescue system. > Usually it is small ("one pack") z/OS system, which can be IPLed as a last > resort. > Update RACF db? Yes, but ...no, because some changes can make the system > un-IPLable. > Test it? Yes, frequently. But... it happen to be not frequent. > A lot of risks related to possible human errors, oversights, etc. > Other mention ZZSA or stand-alone ADRDSSU&ICKDSF. > > However we have LPARs. > I'm pretty sure everyone is running a machine with LPARs. (not talking about > museum, zPDT or Hercules) So, what's the problem with keeping some tech > LPAR up and running? > It does consume very few CPU cycles, few memory, same disk space. > The only possible case when the tech system would not help is Power On > Reset. Which is quite rare. Rare enough to prepare for that by testing tech > system - re-IPL before POR. > And of course when you have more than one CPC (DR, sysplex, etc.) even POR > will not shut down some your tech system. > Besides the risk of IPL failure, the advantages of having up and ready z/OS is > quite obvious IMHO. > I use such LPAR since times of OS/390. And yes, I fixed a lot of small and few > big errors during IPL of prod systems. > > -- > Radoslaw Skorupka > Lodz, Poland > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, send email to > [email protected] with the message: INFO IBM-MAIN
---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
