maria-discuss team mailing list archive

[Gordan Bobic <gordan.bobic@xxxxxxxxx>]

On Thu, Dec 8, 2022 at 7:28 PM Reindl Harald <h.reindl@xxxxxxxxxxxxx> wrote:

> >>>> MariaDB Server 10.4 introduced a new file format
> >>>> innodb_checksum_algorithm=full_crc32, and MariaDB Server 10.5 made it
> >>>> the default. Any files that were created when that setting is active
> >>>> are guaranteed to write any unused bytes as zeroes. It also fixes a
> >>>> peculiar design decision that some bytes of the page are not covered
> >>>> by any checksum, and that a page is considered valid if any of the
> >>>> non-full_crc32 checksums happen to produce a match. This includes the
> >>>> magic 0xdeadbeef for innodb_checksum_algorithm=none.
> >>>>
> >>>> Maybe we should consider eventually deprecating write support for the
> >>>> non-full_crc32 format, to force a fresh start.
> >>>
> >>> Please don't. Some of us run MariaDB on file systems that do their own
> >>> block checksumming, and thus run innodb_checksum_algorithm=none
> >>
> >> that's nonsense - when mariadb writes wrong data in it's files no
> >> filesystem can magically fix that
> >
> > MariaDB can't fix it either. And if that is what happened, there is no
> > benefit to duplicating the effort.
>
> MariaDB does the same as the filesystem
> InnoDB in fact is more ore less a FS on top of a FS

So why do it at both levels? And what makes doing it at MariaDB level
in any way better than doing it somewhere else?

> >> you need to understand what innodb checksums are and then it's logical
> >> that the file-system layer is a completly different world
> >>
> >> https://dba.stackexchange.com/questions/171708/what-is-an-innodb-page-checksum
> >
> > You need to understand that properly thought out and sensibly written
> > file systems (which is, granted, pretty rare, I know of a total of 1)
> > implicitly prevent torn pages from being possible.
> > So the checksum and the doublewrite are completely redundant in such
> > cases, and can be safely disabled.
>
> "Some of us run MariaDB on file systems that do their own block
> checksumming, and thus run innodb_checksum_algorithm=none" makes you
> looking like a fool - period
>
> are you dumb or why don't you understand that the filesystem is a
> completly different layer and has no clue about the data itself?

Are you too dumb to understand that if a block is corrupted at InnoDB
level MariaDB can't do anyting to fix it, but if a block is corrupted
at lower level, ZFS can fix it from redundantly stored data and
MariaDB never gets to ingest a corrupted block in the first place?
If you disagree, please describe a scenario in which an InnoDB page
checksum does anything useful if the file system it is on has built in
block checksumming and data redundancy.