on Fri, Dec 19, 2003 at 01:07:47AM -0800, Alvin Oga (aoga@ns.Linux-Consulting.com) wrote:
>
>
> On Fri, 19 Dec 2003, Lucas Albers wrote:
>
> > >
> > > hi ya andrew
> > > raid can break due to:
> > > - (1) disk failures
> > > - the silly system takes forever ( dayz ) to resync itself
> > > - too many disks failures renders the entire raid useless
> > > or the system can be on a non-raided disk and raid5 for data only
> > > - have an 2nd system disk for backup and go live by
> > > simply changing its ip# and hostname
> > > there is no point to raiding /tmp ...
> > > - if the system dies ... all temp data in /tmp wont matter
> > >
> > > - swap is already "semi-raided" by the kernel
> > > and if it dies... swap data is generally useless anyway
> > >
> > > c ya
> > > alvin
> > I was thinking about this idea, so /tmp is on raid. Now temp dies, and you
> > reboot, and now apache won't start?
>
>
> if /tmp is a separate partition and it cannot mount it during bootup,
> nothing will work right if the app depends on /tmp, not just apache
Wrong.
The mount point directory will exist, and will serve as /tmp. The
normally mounted filesystem won't be there. You'll just be using your
root FS as /tmp.
RAIDing /tmp is generally pretty silly in any regard -- you don't need
data reliability, and generally take a performance hit doing this.
*Striping* /tmp might be called for. You can park any arbitrary
filesystem under /tmp to squeeze by in a pinch anyway.
Peace.
--
Karsten M. Self <kmself@ix.netcom.com> http://kmself.home.netcom.com/
What Part of "Gestalt" don't you understand?
Backgrounder on the Caldera/SCO vs. IBM and Linux dispute.
http://sco.iwethey.org/
Attachment:
pgp8rlDgwryL4.pgp
Description: PGP signature