On Tue, 6 Jun 2006 14:13:01 -0400 Jason Lunz wrote: > On Tue, Jun 06, 2006 at 08:53:45AM -0700, dean gaudet wrote:
> > this is why "-Ac partitions" will work... Neil documented this in Thanks alot for the reassuring details. I am convinced now! > I can certainly live with an etch-only solution. yaird's not even in > sarge. I cannot: This is relevant, even for official Debian packaging of yaird - see below! > The downside of '-Ac partitions' is that if used in yaird, then > backported, it'll crash the mdadm in sarge. That's a pretty minor > problem, IMO. I've cc:ed Norbert Tretkowski so he'll be aware that > future yaird backports might depend on an mdadm backport. It is wrong to assume a purely no-older-than-etch system: Imagine the process of upgrading from sarge to etch; for some reason yaird gets installed first, pulled in by a kernel then installed and using yaird, and only later in the the upgrade process mdadm upgraded. The initramfs will then contain a too old mdadm :-( Yaird packaged officially for Debian can be made to either require a new enough mdadm or conflict with too old ones. but this is ugly, as yaird does not really depend on mdadm, and a conflict would potentially throw out mdadm properly working for non-rootfs. A better approach would be for yaird to check at initramfs build time if mdadm supports -Ac partitions. Can someone come up with a test command that does not mess with actual devices, and returns true for versions of mdadm that properly supports "-Ac partitions"? Or even better, a complete patch for yaird? - Jonas -- * Jonas Smedegaard - idealist og Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ - Enden er nær: http://www.shibumi.org/eoti.htm
pgpiBIFbH4SeG.pgp
Description: PGP signature
