Михаил Кечинов wrote:
> Version 8.4
> So, it was not database crash - HDD died. We copied data to new HDD,
> droped some dead indexes (when vacuuming we has errors with indexes,
> so we drop it and recreate new indexes), made vacuum full. That's all.
where did you copy this data from if the drive died? if it was from a
previously made file system backup, was this backup made with some sort
of point-in-time volume snapshot capability, or was pg_start_backup()
called before it was made ?