Jon Buller wrote:
> Isn't "REQUIRE: quota" a bit late in the process to be "before a
> fsck(8)"? Would it be bad to make the egrep output read like this
> instead:
> 
>      raidframe:# PROVIDE: raidframeconfig
>      raidframeparity:# REQUIRE: raidframeconfig
>      raidframeparity:# PROVIDE: disks
> 
> (i.e. move the parity stuff way up in the startup list.  It does
> all run in the background anyway, although that might cause some
> disk seek contention with fsck if they both run at the same time...)

This is precisely the reason why the parity rewrite has been
moved. There is no good reason to start the parity rewrite sooner;
it would only drastically slow down fsck and other bootup operation.
 
> Perhaps related to this is that every time I boot my machine, I
> get a couple of messages late in the process:
> 
>      raidctl: unable to open device file: raid0
>      raidctl: unable to open device file: raid1

You SHOULD NOT use 'c' or 'd' parts of the disk. 'c' is reserved
for 'NetBSD part of the disk' and 'd' for 'whole disk' (on i386).
Create a separate disk part, can use one of a,b,e-p for this.

> However, I can work around that with "raidctl -s /dev/rraid0c" when
> needed.  Is this normal behaviour, a bug, or a broken setup?

Broken setup.

Jaromir 
-- 
Jaromir Dolecek <jdolecek@NetBSD.org>            http://www.NetBSD.org/
-=- We should be mindful of the potential goal, but as the tantric    -=-
-=- Buddhist masters say, ``You may notice during meditation that you -=-
-=- sometimes levitate or glow.   Do not let this distract you.''     -=-