3850 continued.
First a little background. We had moved 6 or 8 years earlier from a three-story building to an 18 story skyscraper (actually it was higher, but we had the bottom 18 floors.) After the first couple of years, the DC started to expand and expand and expand. We went from 1 floor to two floors, and a tape library was out of control as were tape drives as well as disk drives, There wasn’t a week that we didn’t get in more tape & disk drives. We also were expending CPU wise I can’t remember all the iterations but IIRC a 145, 148, 155,158, 168, 168MP, 3033. You get the idea we were exploding growth wise. Our VP was a former IBMer and always got the latest greatest IBM hardware, I am talking serial numbers 5-10 of the unit. At the time of the 168MP, we were expanding disk dramatically, and management divided up into three departments, CICS, MVS, DISK/Tape. At about that Time IBM came up with the 3850 MSS. It was sort of a tape drive that changed the data into a disk drive (behind the scenes) the “tape” was put in a bullet shipped cartridge(s) and stored in the 3850 in a beehive like structure. When the MVS sent down a request for a dataset, it was translated into a request to take one(actually there were two carts per drive) of the cartridges and automatically stage it onto a disk drive. Once some of the data was staged the 3850 did the same thing for each request. The amount of staged data was limited to the amount of staging drives and capacity of the drives. If the data was read only the destining of the data wasn’t needed and was tossed, the changed data had to be written back to the cartridge. The thing worked reasonably well for a couple of months until we started to see many jobs active but not running. They were waiting on the allocation quench’s called Q4. Over time an allocation was requested it had to go through the allocation q4 this caused havoc with jobs that were put into wait state as the Q4 allocation was taking long and longer to get through as a request for a disk drive to be mounted had to go through Q4. What was bad if a job went through abnormal termination and address space failure, the request wasn’t taken off the Q4. Jobs like ONLINEs weren’t affected just regular jobs, and TSO users caused the Q4 to lengthen and lengthen and then just go dead. We were taking standalone 2-3 a day, and 98 percent had to die with allocation. We had IBMers stacked up going through stand-alone dumps. IBM finally figured out that Allocation had to be rewritten. While we were waiting for that to happen more dumps. IBM finally got done and sent out a replacement, and I am thin on the details as to whether it was via PTF or a release of the OS (sorry its been *AGES*). I do remember one evening I was working late and the console phone rand and the operator called out a coworkers name, and I said he wasn’t here this week, so then the person on the phone asked for the guy's co-worker and I said he went home at 5 PM (it was 7 PM). I went over to the phone and found out that it was IBM in Boulder Colorado. I asked if I could help and they asked me if I was 3850 trained and I said no but knew enough of the concepts that maybe I could help. He asked me if I knew how to do a table zap. A table in MSS terms was the brains of the MSS (so to speak) and it was where the MSS kept track of what was on the live disk drives. I said sure I had seen it done plenty of times, so I fired the STC up and he asked me to do some vers and I read off what was the results and after a few minutes he wanted to do some reps, I was a little nervous and asked are you sure? They said yes. So I proceeded to do the reps. I noticed that quite a few jobs ended and the system seemed to be normal. The two trained 3850 people were not around, so I became the expert of the minute. I wrote up a paragraph on what I did and who I talked with. The next day the 3850 experts (ours) tried two raise a ruckus at what I did, I asked him do we want to go there as I had covered for him and his juniors job that evening. He backed off. His cohort was, at it turned out, that didn’t let work interfere with his social life. He left his desk at the same time every day and came at the same time every day, not conducive to IT work. He (the jr) was pissed at me ever since. He held it against me when we turned up at a different company. But he never talked to me again unless he needed information and I was the only person that had it. I never did find out why he was mad but since weren’t on the best of terms I let it go. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
