background_fsck=no does not work?

[Matthias Buelow]

Scott Long <scottl at samsco.org> writes:

>Whether or not its algoritms are correct or the VM and VFS layers
>properly support it, modern IDE write caches pretty much make write
>orderings a crapshoot.

Well.. that can be countered somewhat by using a UPS for servers, or in
areas where power is flakey.  I'm more concerned about the software
being correct, since the handful or so of filesystems I've lost over the
years have always been due to software failure (usually some major
f*ckup following a kernel panic that fsck couldn't repair...)

mkb.