> > First, as I've already mentioned in answer to Tom about DROP TABLE, undo
> > logic will not be implemented in 7.1 -:( Doable for tables but for
indices we
> > would need either in compensation records or in xmin/cmin in index
tuples.
> > So, we'll still live with dust from aborted xactions in our
tables/indices.
>
> Does it mean that there would still be inconsistency between
> tables and their indexes ?
Not related. I just meant to say that tuples inserted into tables/indices by
aborted transactions will stay there till vacuum.
Redo should guarantee that index tuples will not be lost in split operation
(what's possible now), but not that an index will have correct structure
after crash - parent page may be unupdated, what could be handled
at run time.
Vadim