pg.ddx.io pgsql-general@postgresql.org mailing list archivehelp / color / mirror / Atom feed
Indexes mysteriously change to LOG 3+ messages / 2 participants [nested] [flat]
* Indexes mysteriously change to LOG @ 2023-01-27 20:23 Rumpi Gravenstein <rgravens@gmail.com> 2023-01-27 20:34 ` Re: Indexes mysteriously change to LOG Adrian Klaver <adrian.klaver@aklaver.com> 0 siblings, 1 reply; 3+ messages in thread From: Rumpi Gravenstein @ 2023-01-27 20:23 UTC (permalink / raw) To: PostgreSQL <pgsql-general@lists.postgresql.org> We are on PostgreSQL 14.5 on x86_64-pc-linux-gnu, compiled by gcc (GCC) 8.5.0 20210514 (Red Hat 8.5.0-10), 64-bitPostgreSQL 14.5 on x86_64-pc-linux-gnu, compiled by gcc (GCC) 8.5.0 20210514 (Red Hat 8.5.0-10), 64-bit We have recently discovered that on some of our partitioned tables indexes that were created as: CREATE UNIQUE INDEX chapter_u01 USING btree (dur_uk, catalog_id) somehow changed to include the ON ONLY option: CREATE UNIQUE INDEX chapter_u01 *ON ONLY *chapter USING btree (dur_uk, catalog_id) There is no SQL issued that explicitly requests this "ON ONLY" option. I am wondering if this is a side-effect of some other activity. Googling and looking through documentation haven't helped. Does anyone have any thoughts on how this might happen? -- Rumpi Gravenstein ^ permalink raw reply [nested|flat] 3+ messages in thread
* Re: Indexes mysteriously change to LOG 2023-01-27 20:23 Indexes mysteriously change to LOG Rumpi Gravenstein <rgravens@gmail.com> @ 2023-01-27 20:34 ` Adrian Klaver <adrian.klaver@aklaver.com> 2023-01-27 20:51 ` Re: Indexes mysteriously change to LOG Rumpi Gravenstein <rgravens@gmail.com> 0 siblings, 1 reply; 3+ messages in thread From: Adrian Klaver @ 2023-01-27 20:34 UTC (permalink / raw) To: Rumpi Gravenstein <rgravens@gmail.com>; PostgreSQL <pgsql-general@lists.postgresql.org> On 1/27/23 12:23, Rumpi Gravenstein wrote: > We are on PostgreSQL 14.5 on x86_64-pc-linux-gnu, compiled by gcc (GCC) > 8.5.0 20210514 (Red Hat 8.5.0-10), 64-bitPostgreSQL 14.5 on > x86_64-pc-linux-gnu, compiled by gcc (GCC) 8.5.0 20210514 (Red Hat > 8.5.0-10), 64-bit > > We have recently discovered that on some of our partitioned tables > indexes that were created as: > > CREATE UNIQUE INDEX chapter_u01 USING btree (dur_uk, catalog_id) How did you do the above without the table name? > > somehow changed to include the ON ONLY option: > > CREATE UNIQUE INDEX chapter_u01 *ON ONLY *chapter USING btree (dur_uk, > catalog_id) > > There is no SQL issued that explicitly requests this "ON ONLY" option. > I am wondering if this is a side-effect of some other activity. > Googling and looking through documentation haven't helped. > > Does anyone have any thoughts on how this might happen? > > -- > Rumpi Gravenstein -- Adrian Klaver adrian.klaver@aklaver.com ^ permalink raw reply [nested|flat] 3+ messages in thread
* Re: Indexes mysteriously change to LOG 2023-01-27 20:23 Indexes mysteriously change to LOG Rumpi Gravenstein <rgravens@gmail.com> 2023-01-27 20:34 ` Re: Indexes mysteriously change to LOG Adrian Klaver <adrian.klaver@aklaver.com> @ 2023-01-27 20:51 ` Rumpi Gravenstein <rgravens@gmail.com> 0 siblings, 0 replies; 3+ messages in thread From: Rumpi Gravenstein @ 2023-01-27 20:51 UTC (permalink / raw) To: Adrian Klaver <adrian.klaver@aklaver.com>; +Cc: PostgreSQL <pgsql-general@lists.postgresql.org> > CREATE UNIQUE INDEX chapter_u01 USING btree (dur_uk, catalog_id) How did you do the above without the table name? That's a cut/paste error. The original index create is: create unique index chapter_u01 on chapter (dur_uk,catalog_id); On Fri, Jan 27, 2023 at 3:34 PM Adrian Klaver <adrian.klaver@aklaver.com> wrote: > On 1/27/23 12:23, Rumpi Gravenstein wrote: > > We are on PostgreSQL 14.5 on x86_64-pc-linux-gnu, compiled by gcc (GCC) > > 8.5.0 20210514 (Red Hat 8.5.0-10), 64-bitPostgreSQL 14.5 on > > x86_64-pc-linux-gnu, compiled by gcc (GCC) 8.5.0 20210514 (Red Hat > > 8.5.0-10), 64-bit > > > > We have recently discovered that on some of our partitioned tables > > indexes that were created as: > > > > CREATE UNIQUE INDEX chapter_u01 USING btree (dur_uk, catalog_id) > > How did you do the above without the table name? > > > > > somehow changed to include the ON ONLY option: > > > > CREATE UNIQUE INDEX chapter_u01 *ON ONLY *chapter USING btree (dur_uk, > > catalog_id) > > > > There is no SQL issued that explicitly requests this "ON ONLY" option. > > I am wondering if this is a side-effect of some other activity. > > Googling and looking through documentation haven't helped. > > > > Does anyone have any thoughts on how this might happen? > > > > -- > > Rumpi Gravenstein > > -- > Adrian Klaver > adrian.klaver@aklaver.com > > -- Rumpi Gravenstein ^ permalink raw reply [nested|flat] 3+ messages in thread
end of thread, other threads:[~2023-01-27 20:51 UTC | newest] Thread overview: 3+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2023-01-27 20:23 Indexes mysteriously change to LOG Rumpi Gravenstein <rgravens@gmail.com> 2023-01-27 20:34 ` Adrian Klaver <adrian.klaver@aklaver.com> 2023-01-27 20:51 ` Rumpi Gravenstein <rgravens@gmail.com>
This inbox is served by DDX for PostgreSQL; see mirroring instructions for how to clone and mirror all data and code used for this inbox