| From: | Michael Paquier <michael(at)paquier(dot)xyz> |
|---|---|
| To: | Andres Freund <andres(at)anarazel(dot)de> |
| Cc: | Robert Treat <rob(at)xzilla(dot)net>, Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com>, Sergei Kornilov <sk(at)zsrv(dot)org>, "pgsql-hackers(at)lists(dot)postgresql(dot)org" <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: REINDEX CONCURRENTLY 2.0 |
| Date: | 2019-03-30 01:44:09 |
| Message-ID: | 20190330014409.GO1954@paquier.xyz |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Fri, Mar 29, 2019 at 08:48:03AM -0700, Andres Freund wrote:
> Yes, it increases the total runtime quite considerably. And it adds new
> failure modes with partially built invalid indexes hanging around that
> need to be dropped manually.
On top of that CONCURRENTLY needs multiple transactions to perform its
different phases for each transaction: build, validation, swap and
cleanup. So it cannot run in a transaction block. Having a separate
option makes the most sense.
> It does at *least* twice as much IO.
Yeah, I can guarantee you that it is much slower, at the advantage of
being lock-free.
--
Michael
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Michael Paquier | 2019-03-30 01:51:23 | Re: clean up pg_checksums.sgml |
| Previous Message | Michael Paquier | 2019-03-30 01:19:17 | Re: REINDEX CONCURRENTLY 2.0 |