help! axp/oracle7 odd behaviour

From: sonny hays-eberts <eberts_at_donald.uoregon.edu>
Date: Wed, 12 Oct 1994 15:24:10 -0700
Message-ID: <eberts-121094152411_at_shays.uoregon.edu>


we're having some wierd problems, and i wonder if anybody else has any ideas or, even better, solutions.

we recently moved from oracle v6 to oracle7, while simultaneously moving from a standard vax to an axp environment. one result is that all the tables and data should be pretty clean.

but, here's the problem. certain tables, (it's not confined to any particular one, but does seem to be related to frequency of access) occaisionally lock up. there are no locks you see in the v$xxx views, but any job that tries to access a particular record goes into a permanent HIB state, and nothing else can access that record. of course, if you use an index, you can retrieve data from the index, but that's not much help... :)

if we bounce the database, it clears up for a day or. we thought the problem might be related to risc pipelining and running jobs at different priorities, but it seems not to be the source. since it's not showing up as a lock, we guess it's some sort of latch problem.

oracle doesn't know, and also doesn't seem to think that the inability to access all our data in a production environment isn't a big problem. feh.

i can supply more details, if needed. any clues? please respond via email.

thanks, -sonny

-- 

don't lean on me man,
cause you can't afford the ticket
   -david bowie

sonny hays-eberts                
eberts_at_donald.uoregon.edu
Received on Wed Oct 12 1994 - 23:24:10 CET

Original text of this message