volume by volume salvage by default. This way, servers with lots of files
can be up salvaging volume by volume instead of down until the last volume
is salvaged.
"if it works, go for it" ckuethe@
o function'ize openafs-setup and let it setup not just the 1st but also
subsequent afs servers, letting admins choose which afs server programs to
run on each, `works for me(tm)'
bump PKGNAME
- s/%SYSCONFDIR%/etc where it makes sense (e.g. /etc/rc.conf.local will
always be /etc/rc.conf.local whatever value ${SYSCONFDIR} is set to)
- use SUBST_CMD
- don't hardcode /usr/local
- precise license marker
- s/definately/definitely/ (from sthen@)
"looks good" sthen@, ok todd@
o no longer require kernel sources since not building lkm; if this changes,
the attic can be raided
o typo in openafs-setup comments from sthen@; thanks!
o obey CFLAGS and CC
o twiddle lwp.h patch to be OpenBSD specific
bump PKGNAME
fail to exec autoconf-2.60
o update README.OpenBSD to include rc.local and rc.shutdown bits
o invade and cleanup openafs-setup
- polish consistency
- prompt for a sample general purpose user
- backup CellServDB and restore on script failure
- compact output
- better cleanup incase it is restarted
- show all afs commands as they are run
- reorder various places to make more logical sense
- DO NOT DO RECURSIVE MOUNTS, this leads to panics;
(thanks bob/art for explaining).. i.e. no 'fs mkm /afs/.root.afs root.afs')
- add a few gratuitous remote cells during setup as examples
- add files of consequence to /etc/changelist
bump PKGNAME
o add README.OpenBSD (includes heimdal setup instructions)
o rework openafs-setup to use arla's afsd in base, and heimdal
o fix a bug since 4.1 to not blow the thread stack on 64bit time functions in libc
