Bug, index problem??

From: Terry Mackintosh <terry(at)terrym(dot)com>
To: PostgreSQL-development <hackers(at)postgreSQL(dot)org>
Subject: Bug, index problem??
Date: 1998-10-02 20:00:39
Message-ID: Pine.LNX.3.95.981002151748.18310B-100000@terry1.acun.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Where to start ... base system:
Red Hat 4.2, flex 2.5.4, 2.0.35 kernel
(3 differant boxes, all Pentium+, all 90MB+ RAM)

pg version:
6.4 snap shot, several over the last few weeks.
(the last install went perfect, passing all regresion tests)

Problem:
A script that does an insert, then about a dozen updates on tiny tables
gets slower and slower each time it runs, till it finally just will not
return.

What I have found out:
It has some thing to do with a unique index the never exited in the first
place.

This may be related to some of the posts that I have been seeing here
lately, but most of them go over my head, so I'm not sure.

Basically, a 2 field table (char() and bool) has a unique index on the
char() field, but there is NO index on the bool field, but if one
repeatedly runs some things like:
update table set boolfield = 'f';
(once each on several similar tables)
then
update table set boolfield = 't' where charfield = 'something';
(once each on several similar tables)

after this is done a few times, some where along the line postgres starts
to thing there is a unique index on the boolean field of one or more of
the tables???

I did have it once telling me that it could not do the secound update,
it said that it can not insert a duplicate value in a unique field!

If any one is interested, I do have some screen snips of psql.
The whole database is very small, I can even email a dump file and
associated PHP 3 scripts to any one that may be that interested.

I have reinstalled several times, with several differant snap shots,
hoping the problem whould go away, but it is still here.
At this point I am going to get the latest snap shoot and try it all from
scratch, trying to make a 'test case'.

I will try to make a simple case that can show the problem.

Thanks all, have a great day
Terry Mackintosh <terry(at)terrym(dot)com> http://www.terrym.com
sysadmin/owner Please! No MIME encoded or HTML mail, unless needed.

Proudly powered by R H Linux 4.2, Apache 1..3, PHP 3, PostgreSQL 6.3
-------------------------------------------------------------------
Success Is A Choice ... book by Rick Patino, get it, read it!

Browse pgsql-hackers by date

  From Date Subject
Next Message teunis 1998-10-02 20:08:23 egcs, glibc-2.0.7, linux - any problems known?
Previous Message Jan Wieck 1998-10-02 19:05:55 Re: [HACKERS] Open 6.4 items