On Thu, 22 Dec 2011 11:46:30 +0100, md@Linux.IT (Marco d'Itri) wrote: > On Dec 22, Philip Hands <phil@hands.com> wrote: > > > It's not as though the proposal doesn't require me to do that > > repartitioning already (if I happen to have such a setup), but > No, it does not. Worst case, it requires you to use an initramfs (of > some kind, it needs not to be the one generated by initramfs-tools). Righto - so I generally do use initramfs anyway, and I've been allocating at least 250MB for /boot for years, so perhaps I've just missed the point with the heat rather than light that's been going on in this discussion. If this "/usr must be mounted early" change goes in, and if I have a system that has / and /boot on /dev/md1 and /dev/md2, with a bunch of other RAIDs stuck together via LVM and then used to provide /usr and the rest of the system, and some idiot in the hosting centre pulls the wrong disk out of an already degraded array, thus confusing it to the point where its not going to start without my help (as happened to me recently): What am I left with? If the answer is that I get to fiddle about with busybox in the initramfs, because most of the tools I'm looking for have been put on the now lost /usr, then I feel that I've lost a significant asset. In that case, my solution would be to quadruple the size of /dev/md2 (i.e. /) in future installs, and put /usr on there too, which sounds like a repartition for existing installs if I want to maintain my current level of robustness. Even then, having /usr on that partition increases the number of writes, and having it bigger increases the chance of it being hit by disk errors, so to some small extent this reduces the resilience of the system. I'm still yet to understand the significant upsides of this proposal, so if you want to persuade the doubters, rather than simply dismiss us out of hand, perhaps you should try spelling them out -- I'm sure from your point of view they are so obvious that you cannot imagine that we all wouldn't know about them already, but since nobody's explicitly stated them, and since there are doubters, I think you can safely assume that some of us are yet to perceive the advantages clearly. Cheers, Phil. -- |)| Philip Hands [+44 (0)20 8530 9560] http://www.hands.com/ |-| HANDS.COM Ltd. http://www.uk.debian.org/ |(| 10 Onslow Gardens, South Woodford, London E18 1NE ENGLAND
Attachment:
pgp41nnsA5UDp.pgp
Description: PGP signature