Greg Stark <gsstark@mit.edu> writes:
> I'm not clear that building from WAL is really going to be that much faster.
> A) algorithmically it's only the factor of log(n) that you're talking about.
> and B) the WAL will have records for every write, not just the final product,
> so it might potentially have a lot more writes to do.
Wrong ... what we log in WAL for a btree index build is just the series
of completed index page images. Recreation of the index would proceed
at whatever your disk read/write bandwidth is.
Like Alvaro, I suspect that people who are using PITR will be concerned
about recovery time, and would not be thrilled with any scenario that
involves REINDEX to get the system back on its feet.
regards, tom lane