Re: NetBSD on a database server - WAPBL, softdeps

[Mindaugas Rasiukevicius <rmind%netbsd.org@localhost>]

"Steven M. Bellovin" <smb%cs.columbia.edu@localhost> wrote:
> The issue Thor is pointing out is the granularity of recovery.  WAPBL
> and fsck use file-level granularity -- they ensure the consistency of
> the *file system*.  Databases have their own, complex internal
> structure; WAPBL and fsck know nothing of them.  If you want
> consistency, you have to ensure that the important blocks are written
> out in the proper order.  It's reasonable to want the database to do
> that; it is not reasonable to expect the system to.

Yes, for example, PostgreSQL has WAL.  Theoretically, database transaction
log does not need an additional file-system journal.  However, my point
was not about the consistency - it was about recovery time.

> I think the system call you should be concerned about is
> fsync_range().  Is that used (properly) by your database?  Is it fast
> enough?

No, AFAIK databases use fsync() (maybe DB2 uses fsync_range?), although
it depends at some level.  You can change WAL sync method in PostgreSQL,
as well as tune data/commit synchronisation behaviour in most databases.

-- 
Mindaugas