Re: min_wal_size > max_wal_size is accepted
От | Marc Rechté |
---|---|
Тема | Re: min_wal_size > max_wal_size is accepted |
Дата | |
Msg-id | 9b2cd7f0-a744-5b10-87e5-b3ebdbbf2d9f@rechte.fr обсуждение исходный текст |
Ответ на | Re: min_wal_size > max_wal_size is accepted (Guillaume Lelarge <guillaume@lelarge.info>) |
Список | pgsql-hackers |
> Hi, > > Le jeu. 7 mai 2020 à 11:13, Marc Rechté <marc4@rechte.fr > <mailto:marc4@rechte.fr>> a écrit : > > Hello, > > It is possible to startup an instance with min > max, without the > system > complaining: > > mrechte=# show min_wal_size ; > > 2020-05-07 11:12:11.422 CEST [11098] LOG: durée : 0.279 ms > > min_wal_size > > -------------- > > 128MB > > (1 ligne) > > > > mrechte=# show max_wal_size ; > > 2020-05-07 11:12:12.814 CEST [11098] LOG: durée : 0.275 ms > > max_wal_size > > -------------- > > 64MB > > (1 ligne) > > > This could be an issue ? > > > I don't see how this could be an issue. You'll get a checkpoint every > time 64MB have been written before checkpoint_timeout kicked in. And WAL > files will be removed if you have more than 128MB of them. > > Not the smartest configuration, but not a damaging one either. > > > -- > Guillaume. I have some doubts when I see such code in backend/access/transam/xlog.c:2334 if (recycleSegNo < minSegNo) recycleSegNo = minSegNo; if (recycleSegNo > maxSegNo) recycleSegNo = maxSegNo;
В списке pgsql-hackers по дате отправления: