Tom Lane wrote:
> Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> > Oh. Is that good? If we set the sequence xid as frozen, how do we
> > rollback a sequence creation?
> If we roll back a sequence creation, the whole table goes away.
> Conversely, even if we commit, no one can see the sequence until they
> can see the pg_class entry for it. Thus, it doesn't matter what's in
> the tuple.
OK, so you are using the pg_class visibility/destuction as an
optimitation to prevent vacuum of sequences --- sounds fine.
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026
In response to
pgsql-committers by date
|Next:||From: momjian||Date: 2002-01-11 20:48:48|
|Subject: pgsql/doc/src/sgml/ref pg_upgrade.sgml|
|Previous:||From: Tom Lane||Date: 2002-01-11 20:44:18|
|Subject: Re: pgsql/src/backend/commands sequence.c |