On 2017-03-05 12:01:07 +0100, Peter J. Holzer wrote:
[...]
> At the current rate of inserts, this threshold will be reached on
> March 24nd. I'll check whether the table is analyzed then.
It was (a little earlier than expected because pg_class.reltuples didn't
increase in the meantime).
On 2017-03-05 08:39:05 -0800, Adrian Klaver wrote:
> On 03/05/2017 03:01 AM, Peter J. Holzer wrote:
> >So it is likely that something happened on that day (disk full?) which
> >wiped out the contents of pg_stat_user_tables.
>
> Are there any logs from that time, either Postgres or system?
>
> I
On 03/05/2017 03:01 AM, Peter J. Holzer wrote:
On 2017-03-03 06:39:35 -0800, Adrian Klaver wrote:
On 03/03/2017 12:33 AM, Peter J. Holzer wrote:
This is with PostgreSQL 9.5.6 on Debian Linux.
I noticed that according to pg_stat_user_tables autoanalyze has never
run on a lot of tables. Here is
On 2017-03-03 06:39:35 -0800, Adrian Klaver wrote:
> On 03/03/2017 12:33 AM, Peter J. Holzer wrote:
> >This is with PostgreSQL 9.5.6 on Debian Linux.
> >
> >I noticed that according to pg_stat_user_tables autoanalyze has never
> >run on a lot of tables. Here is one example:
> >
> >wdsah=> select *
On 03/03/2017 12:33 AM, Peter J. Holzer wrote:
This is with PostgreSQL 9.5.6 on Debian Linux.
I noticed that according to pg_stat_user_tables autoanalyze has never
run on a lot of tables. Here is one example:
wdsah=> select * from pg_stat_user_tables where schemaname='public' and
This is with PostgreSQL 9.5.6 on Debian Linux.
I noticed that according to pg_stat_user_tables autoanalyze has never
run on a lot of tables. Here is one example:
wdsah=> select * from pg_stat_user_tables where schemaname='public' and
relname='facttable_wds_indexstats';
─[ RECORD 1