RE: Unique indexes on system tables

> -----Original Message-----
> From: Bruce Momjian [mailto:pgman(at)candle(dot)pha(dot)pa(dot)us]
> Sent: Tuesday, November 16, 1999 1:38 PM
> To: Hiroshi Inoue
> Cc: PostgreSQL-development
> Subject: Unique indexes on system tables
>
>
> Hiroshi, there are two things I want to do for 7.0.
>
> First, I want to make more of the system indexes unique. Are you aware
> of any reasons not to do that?

No.
It should be done to guarantee the uniqueness of system tuples.

> I see you have done some of them
> already. I talked to Tom Lane, and he thinks that it will not cause
> problems because unique insertions/updates wait for transactions to
> commit before doing a conflicting change to the index, right?
>

Yes.

> Second, I want to add more system indexes to match all caches. Anything
> that could cause problems there?
>

I am only afraid of index corruption.
The more we have system indexes,the more index corruption would happen.

How could we recover from the state ?
Tom suggested rebuilding indexes in vacuum.
According to Jan,there was a utility called reindexdb.
WAL by Vadim may be able to recover indexes completely in case of crash.

> Are you working on any of this?
>

No.

Regards.

Hiroshi Inoue
Inoue(at)tpf(dot)co(dot)jp