HP3000-L Archives

June 1997, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Reply To:
Date:
Mon, 9 Jun 1997 20:32:04 -0400
Content-Type:
text/plain
Parts/Attachments:
text/plain (34 lines)
I have encountered a mysterious problem.
One day, two data bases, out of many on a 960, had indexing disabled,
rollback recovery enabled, and logging disabled.  No other changes
occured.  We are unable to find any job that could have done this on
purpose, or any other explanation for how this might have happened.
First question: any ideas on what to look for as a cause?

One of the data bases was the OCS Express data base (for job
scheduling), OCSDB.  OCSDB has Superdex indexes, and they are
used by programs.  When it was discoverd that indexing had been
disabled, entries had been made to OCSDB, so it was necessary to
re-index.  The re-indexing job ran with no errors and all the correct
messages and record counts.  SIMAINT says the indexes are there,
just as they were before, and ADAGER finds nothing wrong with the
data base.  However, the programs that use the indexes do not
select any records, even for values that we can see in the data base.
QUERY does not find any records for any values we try either.

OCSDB also exists on the other two HP3000's, with the same indices.
QUERY and the same programs work as usual on those two machines.
All three machines are on MPE/iX 5.5, PowerPatch 1, same versions
of SUPERDEX, QUERY, and ADAGER.  I copied OCSDB from one of
the machines where it works to the machine where it doesn't work, and
QUERY and the programs work as usual on this copy.  We cannot find
any difference between the two versions of OCSDB, other than the data
itself.

Bradmark, Adager, and HP can suggest nothing other than looking for a
job that might have changed the flags, but no one has any idea about
how to fix the non-functioning indices.  ANY suggestions would be
gratefully received!

Cecile Chi   ([log in to unmask])

ATOM RSS1 RSS2