HP3000-L Archives

March 1995, Week 4

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:
Eric Schubert <[log in to unmask]>
Reply To:
Eric Schubert <[log in to unmask]>
Date:
Fri, 24 Mar 1995 11:45:37 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (30 lines)
At 09:10 AM 3/24/95 MST, Rich Trapp wrote:
>Eric Schubert writes concerning Omnidex:
>> I have always been a staunch critic of how Omnidex handled Index updates (we
>> do Image logging - it's not pretty) and the administrative work involved to
>> keep this thing alive [my time :-( ].
>
>Did you ever try external indexing with the 2.10 version?  This keeps our index
>datasets in a separate database and then you could log only your data?  The
>overhead is 1 more dbopen and MR capability required on programs.
 
The external indexing had bad implications for the logging cycle and logical
transaction support.  I wasn't about to risk aborts from DBRECOV (then a CM
program) for the trade off or spend 4-6 hours rebuilding indexes (the OTHER
problem here - SIZE).  Not too many people actually understand what the
problems are - using Image logging DBEGINS and DBENDS to backout LOGICAL
transactions.  But I can tell you, our people don't panic on system aborts
and fix data with query in a frenzy when it happens!  We never had a broken
chain in 12 years of production (since 1983), how many sites can make that
claim?!
 
The concept of a logical transaction seems to elude many on 3k's - since
only about 2% of the HP sites use these methods.  We are so damn academic
here we actually code logical transactions and do logging by the book!
--------------------------------------------------------------------
Eric J. Schubert                 Administrative Information Services
Senior Data Base Analyst         University of Notre Dame, IN USA
 
Email:            [log in to unmask]
World Wide Web:   http://www.nd.edu/~eschuber

ATOM RSS1 RSS2