EMC 64K chip memory

[UseNet News]

On June 17, we had two of our vax 780s (seismo and hugo) upgraded from
4Mbytes of old style memory to 16 Mbytes of new memory.
These were the 4Mbyte boards with 256k chips, not the 1 M boards.

While our experience was mixed, we are satisfied and would do it again.

The installation on hugo was first. It took about 2 hours from halting
the system to coming back up in production. It went flawlessly. Unix
found the full 16 Mbytes, etc (This was 4.3BSD beta).

Then we started on seismo. They discovered that they had two bad backplanes
(the one for seismo and their spare). A different person came back a week
later to try again with new backplanes.

The installation of seismo took about SIX hours. Two to install the
memory and FOUR hours of me and the CE arguing over whether Unix
was broken since it only found 8 of the 16 M bytes. He claimed his
diagnostics found the full 16Mbytes. Unix only found 8.

The CE was your basic "I only understand VMS" type. However, his
supervisor, who I talked to on the phone was a complete asshole. Had
he been in the room with me, I probably would have hit him. I can understand the
CE's not being familiar with Unix, but their supervisor should have some
experience.

The supervisor insisted that Unix could not see more than 8 Mbytes
without modifications and that I obviously had not installed the modifications
they had given me (These were for System 5 and totally useless). He also seemed
unable to understand that there was more than 1 kind of Unix. I pointed out
that the identical kernel running on the Vax they upgraded last week had
happily found the full 16 M and I didn't believe it was a software problem.
He didn't think the comparison was relevant. (He never did explain why)

While we were arguing on the phone, the CE came back in and said that
he rebooted the system and now Unix found the full 16Mbytes. He denied
making any hardware changes, but our operator saw him fooling with some 
boards. I know the kernel hadn't changed.

Anyway, we finally got the system up and both machines have been running
fine ever since. I am still extremely pissed off at the EMC manager
who I spoke to on the phone. I haven't run into that kind of anti-Unix
paranoia in about 5 years (Remember what DEC field service was like
when the Vax first came out? You pretty much had to make it fail under
VMS before they would believe it was a problem). This cretin is giving
EMC an undeservedly bad reputation. 

So, the short answer on EMC is that had we only had one system upgrade, they
would have been perfect. But, based on the other one, they need some
work in customer relations. Someone less experienced with Vaxes and Unix
might have been intimidated into accepting a broken system.

---rick