On Tue, Oct 19, 2004 at 03:04:54PM +0900, Rob wrote: > Ryan Sommers wrote: > >Jose M Rodriguez wrote: > > > >>If You try to path /etc/rc.d/initdiskless, try not remove *.cpio.gz > >> > >>expand using /stand, but add *.tar expand using /bin/pax. > >> > >>You may also try some auto detection and use what is present between > >>/stand/cpio, /stand/gzip, /rescue/tar, /rescue/gzip ... > >> > >>Don't try moving gzip/libgzip from /usr to rootfs. It's really a big > >>change. > >> > >/rescue/* aren't the right tools for the job in my opinion. Neither are > >/stand/*. I posted this to arch_at_ [1] and between that and this thread, > >among others on current_at_, the overwhelming majority of responses have > >been to get rid of /stand. I'm not sure I understand your argument for > >leaving /stand as part of the root filesystem; other than as you say > >your 250gb drives don't mind the extra 3.5mb. In my opinion it's useless > >for both diskless and regular systems, with the exception of 2% of it's > >space. It's not updated as part of world, which means no bug fixes get > >populated to it. It's a crunchgen binary which means trying to fix small > >bugs in single programs either means rebuilding the binary or replacing > >that hardlink with an individual binary. > > Sorry, I have changed my mind concerning patching my 5.3 system wrt. > this /stand vs. /rescue issue. I'd rather wait until this discussion > is resolved and moreover, until 5.3 has become STABLE and 6.0 has > become CURRENT. > > For now, I will populate the /conf/base directory with the plain files > (not using zipped cpio-archives), in which case I do not need /stand. > > The disk space used is then: > 324K using etc and var as cpio.gz archives in /conf/base > 1.8M using etc and var as plain files in /conf/base > 2.3M /stand > > *.cpio.gz + /stand = 2.6M > plain files - /stand = 1.8M > > For now, with 5.3, I'll go for the plain files without /stand. > I can wait until 6-Stable for this to get reduced to 324K :). You also may want to remember that the plain files have one hugh advantage, they can be modified directly rather then having to rebuild the tarball every time. Most of the time, I can't see how the cpio archives would be worth the trouble. -- Brooks -- Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:18 UTC