summaryrefslogtreecommitdiff
path: root/distrib/miniroot/install.sh
AgeCommit message (Collapse)Author
2014-03-20FTP is so 20th century. Remove 'ftp' as a method to obtain installationKenneth R Westerback
sets. 'http' is the present. A bunch of related or triggered cleanup/consistenizing. Prodded by deraadt@, much feedback and testing by rpe@. ok rpe@
2014-02-21remove stupid comment from decades ago. Proof is in the pudding, turnsTheo de Raadt
out ksh is the perfect language to write this in.
2014-02-21oops, seperate pieces coming...Theo de Raadt
2014-02-21servicesTheo de Raadt
2014-02-04simplify removal of fstab leftovers in /tmpRobert Peichaer
OK halex@
2014-02-02remove -e sed option where only one command is usedRobert Peichaer
ok halex@ krw@
2014-01-05Like upgrade, use feed_random right after mounting filesystems. ThisTheo de Raadt
is less than ideal. We hope the kernel has good entropy already, because this is a reinstall?? Or maybe the ftp server startup has run to completion? Or if this is an autoinstall, that config has enough unique in it? This needs more work, but the existing late-entropy feed sequence is worse. discussed at length with rpe, who tested it
2013-12-18change password prompts (for the good of install.conf)Alexander Hall
add autoinstall question for root ssh pubkey make pubkey prompts appear in autoinstall log ideas from and ok deraaddt@, ok rpe@
2013-12-06Workaround closed stdin in non-interactive mode.Robert Peichaer
ok deraadt@
2013-12-01Allows the autoinstaller to pass a public ssh key for inclusion in theAlexander Hall
user's .ssh/authorized_keys. Deliberately not in interactive mode since I am not sure it's worth adding another question that only quite few people would enjoy. ok rpe@
2013-11-19Add autoinstall configuration to the list of files and output stirringAlexander Hall
the random pool at install. Apart from the configuration in itself, this permits adding various random data in comments or whatever. ok krw@ rpe@
2013-11-16Remove AUTOROOT variable which is a leftover of r1.183 of install.subRobert Peichaer
diff from Philipp e1c1bac6253dc54a1e89ddc046585792 at osteo dot net ok krw@ halex@
2013-10-31If if looks like a duck, swims like a duck, and quacks like a duck, thenAlexander Hall
it's probably is a pre-encrypted password hash. This means that the autoinstall configuration (and interactive password too) does not have to specify a cleartext password. reworked diff originating from krw@ no objections, specifically so from krw@
2013-05-31remove unneeded '\' for line continuationsRobert Peichaer
ok halex@ krw@
2013-04-15Use the dd and ed that are on the media, instead of the one post-install.Theo de Raadt
Amazing this mistake lurked for so long...
2012-09-28consistently use [[ ]] for testsRobert Peichaer
ok halex@
2012-09-28remove unnecessary quotes within [[ ]]Robert Peichaer
ok halex@
2012-09-28replace the last remaining backticks with $() for consistencyRobert Peichaer
ok halex@
2012-09-28remove a comment that does not make sense any moreRobert Peichaer
ok halex@
2012-09-28replace _fsent array with a simple concatenated listRobert Peichaer
ok halex@
2012-09-03compress arguments to the ftp command; ok halexTheo de Raadt
2012-01-31redirect stdout to /dev/null instead of closing itAlexander Hall
ok krw@
2011-08-17make sure to set two rtsol sysctls, not just the last oneTodd T. Fries
ok krw@
2011-07-24use the right sysctl, doh!Federico G. Schwindt
pointed out by Moritz Grimm (mgrimm-at-mrsserver-dot-net)
2011-07-23uncomment net.inet6.ip6.rediraccept from /etc/sysctl.conf as well if theFederico G. Schwindt
user picks up rtsol. reported in bugs@ by Rene Maroufi (info at maroufi dot net). bluhm@ ok
2011-07-04there is no need to do the rtsol check (whether to enableAlexander Hall
net.inet6.ip6.accept_rtadv in sysctl.conf or not) on updates as then we dont do any network config, so move it from install.sub to install.sh idea not rejected by IPv6-enabled bluhm@ ok krw@ deraadt@
2011-04-17Make DUID fstab the default so we can see how people like it. RequestedKenneth R Westerback
by deraadt@.
2011-04-17First crack at enabling the installation of a DUID version ofKenneth R Westerback
/etc/fstab, after asking user. Current default is existing behaviour. Feedback & suggestions deraadt@, halex@, jsing@, todd@. ok deraadt@
2011-04-04If we choose to create a user during installation, put it intoAntoine Jacoutot
/root/.forward. "makes sense" kettenis, ok pirofti@ deraadt@
2011-02-07Use absolute path to the temporary hosts file; causes /etc/hosts to be correctlyMiod Vallat
populated upon installation again.
2011-01-10like in /etc/rc, talk to /dev/arandom as a single read or write.Theo de Raadt
in particular a single write will result in a single re-key event, rather than 64 writes causing 64 re-keys -- wasting the kernel's time.
2011-01-03reduce the amount of cd'ing in the install script, since it sucksTheo de Raadt
to change the environment that functions assume they run in. use sub-shells or absolute paths when possible. ok krw
2010-12-22remove redundant junkTodd T. Fries
pointed out by Mikalaj Kucharski, mikolaj at kuharski dot name ok deraadt@
2010-12-22add more commands of output of random "junk" to arandomTodd T. Fries
please commit deraadt@
2010-12-22wow, we must reach into the chroot for /dev/arandom, meaning some ofTheo de Raadt
this code has been busted for quite a while. pointed out by miod
2010-12-22Consistantly use /dev/arandom for feeding entropy to the kernel.Theo de Raadt
ok kjell otto miod
2010-10-30for the user, create a matching group and put the user in there by default; ↵Theo de Raadt
ok halex guenther
2010-10-29for late TZ handling, place the temporarily tzlist file into /mnt/tmpTheo de Raadt
(which has already been mounted). discussed with halex
2010-08-04Set an appropriate To: header in /var/mail/$user which is copied fromAlexander Hall
/var/mail/root idea and first version from deraadt@, ok deraadt@ krw@
2010-04-06Try a more detailed message about the timeTheo de Raadt
2010-04-04If the system time is off by more than 120 seconds from the TIME=nnnAlexander Hall
supplied from ftplist.cgi, ask if the user wants to set it accordingly. Idea from deraadt@, feedback from sthen@, guenther@ ok deraadt@, krw@ (slightly different version)
2009-07-10We should be using the $SERVERLISTALL (not the url-only $SERVERLIST) toTheo de Raadt
decide to talk-back. Sigh.
2009-07-02Do *not* newfs non-ffs partitions while installing. Even if someoneKenneth R Westerback
has entered 'n i' in disklabel and caused a mountpoint to be saved for the non-ffs partition. Discovered by and fix tested by mgrimm@. ok deraadt@.
2009-06-27Add initial user to the 'staff' class.Antoine Jacoutot
ok deraadt@, ok krw@
2009-06-11No need to show the filename the duplicate mount point was detectedKenneth R Westerback
in. Noticed in an error message from todd@.
2009-06-09Update the user input routines to (re-evaluate and) redraw theAlexander Hall
question if dmesg changes are detected. The password reading routines are not subject to these changes at this point. ok deraadt@, krw@
2009-06-03Only need to umount once on restarting. Force umounts in case ofKenneth R Westerback
restarting. No need to ask for verbose umount output since we send it all to /dev/null. Make sure we start in '/', lest we were restarted while exploring the installed/upgraded filesystems.
2009-06-02Nuke sissy verbiage about losing data. No more special casing forKenneth R Westerback
restarted installs/upgrades. They should just work the same as first attempts except for presenting some remembered info as default answers. ok deraadt@
2009-06-01Test the return value of md_prep_disklabel() to detect failedKenneth R Westerback
attempts to write a disklabel. When failure is detected don't configure the disk, return it to the list of disks available to be configured.
2009-05-31push the dmesg and $SERVERLIST into the random pool to enhance theTheo de Raadt
entropy, before we pull out the host.random for the next boot