JOHNSON@NUHUB.ACS.NORTHEASTERN.EDU.UUCP (03/27/87)
Hi folks.
I'm an unfortunate maintainer of RTI's INGRES. Version 3 of INGRES
worked as well as can be expected from people who forget about
universities in their design efforts. Then I got version 5 (please
don't ask why I didn't get version 4, it's a long story).
Up until this time INGRES had always worked and lived on a Fuji
disk off an SI controller. Everything worked fine. Version 5 SEEMED to
work for a time then it broke. I tried everything I could think of from
past failure modes. Nothing worked so I called Tech Support at RTI.
One month later, after much hounding on my part, the answer was
given from on high,
RTI, "Controller bug".
Me, "Uh ... ok.. why did v 3 work?"
RTI, "Don't know. Move INGRES to an HSC based disk."
Me, "But, but, but what's the problem?"
The error I had been getting was something about the device not
supporting byte-aligned transfers of data. It was explained to me that
my Eagles couldn't support this. Not much help as I knew this from the
error message.
So I moved INGRES to an HSC based RA81 and it worked. HOWEVER, I
have my disks set up a certain way for a reason, performance mostly. It
is EXTREMELY inconvenient to move INGRES to an HSC based disk and I don't
have room to keep it there. It has to go back to the Fuji Eagle soon.
They gave me an answer but it happens to be VERY unsatisfactory as far
as I'm concerned.
Anybody out there notice any problems with INGRES on Eagles with an
SI 9900 unibus controller? Anybody fix it? If so, HOW? HELP!
In case the above isn't possible, does anyone have 1032 from
Software House? I'd like to hear any facts/inferences/impressions
anyone has about it especially in a university setting.
Thanks Much
Chris Johnson
Northeastern University
csnet: johnson@nuhub.acs.northeastern.edu
arpa: johnson%nuhub.acs.northeastern.edu@relay.cs.net
at&t: (617) 437-2335