From: Stephen Frost <sfr...@snowman.net> > I hadn't intended to make an argument that enabling checksums was > equivilant to enabling or disabling fsync- I said it was 'akin', by which I > meant it > was similar in character, as in, as I said previously, a way for PG to hedge > against certain external-to-PG risks (though, unfortunately, our checksums > aren't able to actually mitigate any of the risks but merely to detect them, > but > there is certainly value in that too).
I only skimmed a few recent mails in this thread, but I've been hoping that the checksum gets on by default. So +1. As the storage layers get complex, including the container and virtual storage, we may increasingly be passed bogus data by software bugs. I always tell my little kid to fasten a seatbelt, although she dislikes it as she can't move well. I've also felt it sad that checksum is not mentioned in the database cluster setup section of the manual [1]. If checksum is turned on by default, I think we can mention that checksum can be disabled for maximum speed in this page. It'd be desirable to lighten the overhead of checksumming, but I have no idea now. IIRC, Oracle and SQL Server also scan the whole page to compute the checksum. Maybe the disadvantage of Postgres is that it has to copy the whole page for fear of concurrent hint bit updates? [1] 19.2. Creating a Database Cluster https://www.postgresql.org/docs/devel/creating-cluster.html Regards Takayuki Tsunakawa
