Hi, On Wed, Mar 12, 2003 at 04:55:00PM +0100, Thomas Hood wrote: > It is now clear to me that the question about creating /mem > containing a fs that is RAM-based-if-possible is orthogonal > to the question about creating a runtime state directory that > is rw, persistent-until-reboot, and network-independent. Largely, yes. I also agree that /mem should be debated on its own merits, one of it being that it solves an early /var/run. > /mem is neither necessary to solve the runtime state problem > nor sufficient to solve it in every case. I don't understand why it wouldn't be sufficient to solve the runtime state problem in every case? AFAICS it does. > In *this* thread I am proposing that /var/run should be the > standard path to our runtime state directory and that it should > be a symlink (both before and after /var is mounted) to another > directory that is rw, persistent-until-reboot, and network- > independent. It's /very/ similar to Mike's older proposal. The messy part about it is that you put something in /var (the 'run' symlink) that you will have to remove before something can be mounted over /var, at least for older or non-linux kernels. Also, you'll have to keep the symlink in the pre-mounted case synchronised to the symlink in the real /var fs. Accessing the hidden /var after something is mounted on /var will be somewhat of a challenge. Cheers, Emile. -- E-Advies - Emile van Bergen emile@e-advies.nl tel. +31 (0)70 3906153 http://www.e-advies.nl
Attachment:
pgpB7FD0m3kil.pgp
Description: PGP signature