From: | wieck(at)debis(dot)com (Jan Wieck) |
---|---|
To: | peter_e(at)gmx(dot)net |
Cc: | Inoue(at)tpf(dot)co(dot)jp, pgman(at)candle(dot)pha(dot)pa(dot)us, pgsql-hackers(at)postgreSQL(dot)org |
Subject: | Re: [HACKERS] RE: Unique indexes on system tables |
Date: | 1999-11-17 12:39:34 |
Message-ID: | m11o4N4-0003kIC@orion.SAPserv.Hamburg.dsh.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
>
> On Tue, 16 Nov 1999, Hiroshi Inoue wrote:
>
> > I am only afraid of index corruption.
> > The more we have system indexes,the more index corruption would happen.
>
> Just a concerned user question: Why does index corruption seem to happen
> so often or is a genuine concern? Wouldn't the next thing be table
> corruption? Or are indices optimized for speed rather than correctness
> because they don't contain important data?
There are more complicated concurrency issues on indices than
for regular tables. That's where the corrupt indices but not
tables come from.
For a user index, this isn't very critical, because a
drop/create index sequence will recover to consistent data.
For system catalog indices, this is a desaster, because you
cannot drop and recreate indices on system tables. At least
we need to tackle this problem by reincarnating reindexdb.
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#========================================= wieck(at)debis(dot)com (Jan Wieck) #
From | Date | Subject | |
---|---|---|---|
Next Message | Jan Wieck | 1999-11-17 12:48:39 | Re: [HACKERS] regression tests |
Previous Message | Jan Wieck | 1999-11-17 11:46:29 | Re: [HACKERS] regression tests |