Dennis Roach relates his very interesting experience: > ... had a 3031. We added the AP to it. ... The problem did not > occur with the AP offline. ... There was an optional EC that > reduced a section of tri-lead by 18 inches. The EC fixed the > problem.
Lynn Wheeler notes: > remember that 158 & 3031 were the same engine. That's a particularly interesting story, Dennis. I had a 370/158 to which I added an APU. It experienced the exact, same symptoms as you described, and the fix was the very same. As I remember, the original length of the tri-lead was about 7 or 8 feet, while its replacement was just over 5 feet. Maybe it was the same wire! What complicated my situation was that it was a third-party box and had extra OEM [outside-the-frame] memory, and both the CPU and then the APU upgrade were originally installed by a third party (not IBM). When it red-lighted [as you described], we just took the APU offline. The third-party installer tried for weeks (literally) to shoot this bug, but finally gave up (they told me they lost more than $30,000 on that job) and THEY called IBM in. IBM, of course, had been following the saga, and kept telling me to place a (billable) service call and they would fix the box. I could not get management to go along with that, since I couldn't quantify the scope of the risk (i.e., the potential cost). When the third-party installer called IBM, however, it was on _their_ dime. Even though the box was being maintained by IBM, it wasn't IBM's problem until it (along with the APU after being installed by the third-party installer) passed all the IBM diagnostic tests, which of course it was failing. IBM went thru all of the records for the box that the third-party installers were supposed to be maintaining (just as if they were IBM CEs, which of course in a previous life they all were), and checked every single wire and part. The 158-UP they gave us had originally been one side of a 370/158-MP, but, for the original install at my site, downgraded first to a UP and then upgraded with the MES so as to be able to have an APU attached instead, at some point in the future (I was planning ahead). Somewhere along the way, this wire, which was supposed to have been changed out (shortened), was documented as being installed (all the right paperwork for it was in those big blue machine configuration and maintenance documentation binders) but it was NOT, in fact, installed. It took IBM about 3 hours to figure this out. They "borrowed" a shorter tri-lead from the OEM installers to fix the problem, called it a day, and left (later, IBM sent them a bill for the service call for 5 CEs for 4 hours; the third-party installer tried to get _me_ to pay it, saying it was an IBM maintenance problem, but I was not that stupid and got a letter from IBM explaining that the EC associated with the AP MES was not, in fact, correctly installed, as the documentation for the box indicated (hence it was neither my problem no IBM's). We were happy to have our new, faster 370/158-AP. But, except for disk drives, that was the last non-IBM vendor management ever let me do any business with. -- WB ---------------------------------------------------------------------- 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
