> > If the master crashes while a transaction that used CREATE TABLE is
>unfinished,
> > both the master and the standby will indefinitely retain identical, stray
>(not
> > referenced by pg_class) files. The catalogs do reference the relfilenode
of
> > each unlogged relation; currently, that relfilenode never exists on a
>standby
> > while that standby is accepting connections. By the time the startup
>process
> > releases the AccessExclusiveLock acquired by the proposed UNLOGGED ->
normal
> > conversion process, that relfilenode needs to be either fully copied or
>unlinked
> > all over again. (Alternately, find some other way to make sure queries
>don't
> > read the half-copied file.) In effect, the problem is that the relfilenode
>is
> > *not* stray, so its final state does need to be well-defined.
>
> Oh, right.
>
> Maybe we should just put in a rule that a server in Hot Standby mode
> won't ever try to read from an unlogged table (right now we count on
> the fact that there will be nothing to read). If we crash before
> copying the whole file, it won't matter, because the catalogs won't
> have been updated, so we'll refuse to look at it anyway. And we have
> to reinitialize on entering normal running anyway, so we can clean it
> up then.
Ok then... I'll try to code the "easy" version first (the wal_level=minimal
case)
and then we'll see....
Leonardo
