X-Git-Url: https://iankelling.org/git/?a=blobdiff_plain;f=distro-end;h=660d556ea2363d9220ab9859c09fc3d2b4403f97;hb=65351382939fa95fb1e05d7d83eb58d27c3c6133;hp=a30963f5d0e1240db3c5eff77dfbbecba08cbe71;hpb=e688da727d2f4620e52a26e77f190c65d6627e7e;p=distro-setup diff --git a/distro-end b/distro-end index a30963f..660d556 100755 --- a/distro-end +++ b/distro-end @@ -1,170 +1,1139 @@ #!/bin/bash -l # Copyright (C) 2016 Ian Kelling -# This program is under GPL v. 3 or later, see -set -eE -o pipefail -trap 'echo "$0:$LINENO:error: \"$BASH_COMMAND\" returned $?"' ERR + +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at + +# http://www.apache.org/licenses/LICENSE-2.0 + +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +errcatch set -x exec &> >(sudo tee -a /var/log/distro-end) echo "$0: $(date): starting now)" +src="${BASH_SOURCE%/*}" + +source $src/pkgs + +# see example of usage to understand. end_msg() { - = local y + local y IFS= read -r -d '' y ||: end_msg_var+="$y" } +spa() { # simple package add + simple_packages+=($@) +} + distro=$(distro-name) pending_reboot=false +sed="sed --follow-symlinks" # template case $distro in esac pup +if isdeb; then + pi aptitude +fi -# universal packages -simple_packages=( - bwm-ng - chromium - duplicity - evince - fdupes - filelight - gdb - gnome-screenshot - mailutils - meld - mpv - nmon - offlineimap - p7zip - paprefs - pavucontrol - pianobar - pidgin - rdiff-backup - slock - smartmontools - squashfs-tools - tcpdump - transmission-remote-gtk - tree - vim -) +simple_packages=(${p3[@]}) + +case $HOSTNAME in + lj|li) : ;; + *) + simple_packages+=(${p4[@]}) + spa $(apt-cache search ruby[.0-9]+-doc| awk '{print $1}') + ;; +esac -spa() { # simple package add - simple_packages+=($@) -} +########### begin section including li ################ + +conflink case $distro in - debian) pi curl ;; + arch) + # ubuntu 14.04 uses b-cron, + # but its not maintained in arch. + # of the ones in the main repos, cronie is only one maintained. + # fcron appears abandoned software. + pi cronie + sgo cronie + ;; + *) : ;; # other distros come with cron. +esac + + + +case $distro in + *) pi at ;;& + arch) sgo atd ;; +esac + + +case $distro in + debian|trisquel|ubuntu) pi curl;; arch) : ;; # fedora: unknown esac case $distro in - arch) pi syncthing ;; - ubuntu|debian) - # google led me here: - # https://apt.syncthing.net/ - curl -s https://syncthing.net/release-key.txt | sudo apt-key add - - s="deb http://apt.syncthing.net/ syncthing release" - if [[ $(cat /etc/apt/sources.list.d/syncthing.list) != $s ]]; then - echo "$s" | s dd of=/etc/apt/sources.list.d/syncthing.list - p update - fi - pi syncthing + # tk for gitk + arch) spa git tk ;; + *) spa git ;; +esac + +case $distro in + arch) spa the_silver_searcher ;; + debian|trisquel|ubuntu) spa silversearcher-ag ;; + # fedora unknown +esac + +case $distro in + debian|trisquel|ubuntu) spa ntp;; + arch) + pi ntp + sgo ntpd ;; + # others unknown esac -# installed via f-droid -# top right, actions, device id -# -# for installing on a remote comp: -# ssh -L 8384:localhost:8384 -N frodo -# went to http://localhost:8384/ -# -# add folder to sync phone, -# staggered file versioning would be my normal choice, but choose -# trash can versioning for sake of space on phone, with -# clean out after 7 days. -# -# did ser syncthing@ian start -# then on phone, add device, hit bar code icon, -# install bar code scanner. -# things with no equivalent in other distros: +# no equivalent in other distros: case $distro in - debian|ubuntu) - # for gui bug reporting - spa python-vte - pi apt-file aptitude - s apt-file update + debian|trisquel|ubuntu) + pi aptitude + if ! dpkg -s apt-file &>/dev/null; then + # this condition is just a speed optimization + pi apt-file + s apt-file update + fi # for debconf-get-selections spa debconf-utils ;; esac -####### misc packages ########### +case $distro in + arch|debian|trisquel|ubuntu) + spa bash-completion + ;; + # others unknown +esac + + + + +# disable motd junk. case $distro in - ubuntu|debian) - spa spacefm-gtk3 ;; - arch) - spa spacefm ;; + debian) + # allows me to pipe with ssh -t, and gets rid of spam + # http://forums.debian.net/viewtopic.php?f=5&t=85822 + # i'd rather disable the service than comment the init file + # this says disabling the service, it will still get restarted + # but this script doesn't do anything on restart, so it should be fine + s dd of=/var/run/motd.dynamic if=/dev/null + # stretch doesn't have initscripts pkg installed by default + if [[ $(debian-codename) == jessie ]]; then + s update-rc.d motd disable + fi + ;; + trisquel|ubuntu) + # this isn't a complete solution. It still shows me when updates are available, + # but it's no big deal. + s t /etc/update-motd.d/10-help-text /etc/update-motd.d/00-header + ;; +esac + +# automatic updates +# reference: +# https://debian-handbook.info/browse/stable/sect.regular-upgrades.html +# /etc/cron.daily/apt calls unattended-upgrades +# /usr/share/doc/unattended-upgrades# cat README.md +# /etc/apt/apt.conf.d/50unattended-upgrades +if isdebian; then + setup-debian-auto-update +fi + +# we've got a few dependencies later on, so install them now. +pi eatmydata; PI_PREFIX=eatmydata +pi "${simple_packages[@]}" +simple_packages=() + + +### begin docker install #### + +if isdeb; then + # https://store.docker.com/editions/community/docker-ce-server-debian?tab=description + pi software-properties-common apt-transport-https + curl -fsSL https://download.docker.com/linux/$(distro-name-compat)/gpg | sudo apt-key add - + sudo add-apt-repository \ + "deb [arch=amd64] https://download.docker.com/linux/$(distro-name-compat) \ + $(debian-codename-compat) \ + stable" + p update + pi docker-ce + sgo docker + # other distros unknown +fi +### end docker install #### + + +### begin certbot install ### +case $distro in + debian) + # note, need python-certbot-nginx for nginx, but it depends on nginx, + # and I'm not installing nginx by default right now. + # note python-certbot-apache is in suggests, but so is a doc package that brought in xorg + if [[ $(debian-codename) == jessie ]]; then + pi -t jessie-backports certbot python-certbot-apache + else + pi certbot python-certbot-apache + fi + ;; + trisquel|ubuntu) + # not packaged in xenial or flidas + pi software-properties-common + # this fails with: + # + # gpg: key 75BCA694: public key "Launchpad PPA for certbot" imported + # gpg: Total number processed: 1 + # gpg: imported: 1 + # gpg: no valid OpenPGP data found. + # Failed to add key. + # + # but it seems to work fine, perhaps it's only failing on the second run. + s add-apt-repository -y ppa:certbot/certbot ||: + p update + pi python-certbot-apache + ;; + # todo: other distros unknown +esac +# make a version of the certbot timer that emails me. +x=/systemd/system/certbot +$sed -r -f - /lib$x.timer <<'EOF' |s dd of=/etc${x}mail.timer +s,^Description.*,\0 mail version, +EOF +$sed -r -f - /lib$x.service <<'EOF' |s dd of=/etc${x}mail.service +s,(ExecStart=)(/usr/bin/certbot),\1/a/bin/log-quiet/sysd-mail-once certbotmail \2 --renew-hook /a/bin/distro-setup/certbot-renew-hook, +EOF +ser daemon-reload +sgo certbotmail.timer +### end certbot install ### + + +# dogcam setup +case $HOSTNAME in + lj|li) + /a/bin/webcam/install-server + ;; + kw) + /a/bin/webcam/install-client + ;; +esac + +# website setup +case $HOSTNAME in + lj|li) + + case $HOSTNAME in + lj) domain=iank.bid; exit 0 ;; + li) domain=iankelling.org ;; + esac + /a/h/setup.sh $domain + /a/h/build.rb + + sudo -E /a/bin/mediawiki-setup/mw-setup-script + + pi-nostart mumble-server + s $sed -ri "s/^ *(serverpassword=).*/\1$(< /a/bin/bash_unpublished/mumble_pass)/" /etc/mumble-server.ini + + # do certificate to avoid warning about unsigned cert, + # which is overkill for my use, but hey, I'm cool, I know + # how to do this. + web-conf apache2 mumble.iankelling.org + s rm -f /etc/apache2/sites-enabled/mumble.iankelling.org + sudo -i <<'EOF' +export RENEWED_LINEAGE=/etc/letsencrypt/live/mumble.iankelling.org +/a/bin/distro-setup/certbot-renew-hook +EOF + + sgo mumble-server + + vpn-server-setup -rd + s tee /etc/openvpn/client-config/mail <<'EOF' +ifconfig-push 10.8.0.4 255.255.255.0 +EOF + + # it\'s strange. docker seems to make the default for forward + # be drop, but then I set it to accept and it\'s stuck that way, + # I dun know why. But, let\'s make sure we can forward anyways. + s DEBIAN_FRONTEND=noninteractive pi iptables-persistent + rm /etc/iptables/rules.v6 + s tee /etc/iptables/rules.v4 <<'EOF' +*filter +-A FORWARD -i tun+ -o eth0 -j ACCEPT +-A FORWARD -i eth0 -o tun+ -j ACCEPT +COMMIT +EOF + + + sudo dd of=/etc/systemd/system/vpnmail.service < + Options +FollowSymLinks +Multiviews +Indexes + AllowOverride None + AuthType basic + AuthName "Authentication Required" + # setup one time, with root:www-data, 640 + AuthUserFile "/etc/caldav-htpasswd" + Require valid-user + +EOF + # nginx version of above would be: + # auth_basic "Not currently available"; + # auth_basic_user_file /etc/nginx/caldav/htpasswd; + + + ########## begin pump.io setup ########## + + # once pump adds a logrotation script, turn off nologger, + # and add + # "logfile": "/var/log/pumpio/pumpio.log", + # + s dd of=/etc/pump.io.json <<'EOF' +{ + "secret": "SECRET_REPLACE_ME", + "driver": "mongodb", + "params": { "dbname": "pumpio" }, + "noweb": false, + "site": "pump.iankelling.org", + "owner": "Ian Kelling", + "ownerURL": "https://iankelling.org/", + "port": 8001, + "urlPort": 443, + "hostname": "pump.iankelling.org", + "nologger": true, + "datadir": "/home/pumpio/pumpdata", + "enableUploads": true, + "debugClient": false, + "disableRegistration": true, + "noCDN": true, + "key": "/home/pumpio/privkey.pem", + "cert": "/home/pumpio/fullchain.pem", + "address": "localhost", + "sockjs": false +} +EOF + s sed -i "s#SECRET_REPLACE_ME#$(cat /p/c/machine_specific/li/pump-secret)#" /etc/pump.io.json + + # stretch node is too old + # https://nodejs.org/en/download/package-manager/ + curl -sL https://deb.nodesource.com/setup_8.x | sudo -E bash - + pi nodejs graphicsmagick mongodb + cd /home/iank + if [[ -e pump.io ]]; then + cd pump.io + git pull + else + git clone https://github.com/pump-io/pump.io.git + cd pump.io + fi + # note: doing this or the npm install pump.io as root had problems. + npm install + npm run build + # normally, next command would be + # s npm install -g odb + # but it\'s this until a bug in pump gets fixed + # https://github.com/pump-io/pump.io/issues/1287 + s npm install -g databank-mongodb@0.19.2 + if ! getent passwd pumpio &>/dev/null; then + s useradd -m -s /bin/false pumpio + fi + sudo -u pumpio mkdir -p /home/pumpio/pumpdata + # for testing browser when only listening to localhost, + # in the pump.io.json, set hostname localhost, urlPort 5233 + #ssh -L 5233:localhost:5233 li + + s mkdir -p /var/log/pumpio/ + s chown pumpio:pumpio /var/log/pumpio/ + + web-conf - apache2 pump.iankelling.org <<'EOF' +# currently a bug in pump that we cant terminate ssl + SSLProxyEngine On + ProxyPreserveHost On + ProxyPass / https://127.0.0.1:8001/ + ProxyPassReverse / https://127.0.0.1:8001/ + # i have sockjs disabled per people suggesting that + # it won\'t work with apache right now. + # not sure if it would work with this, + # but afaik, this is pointless atm. + + ProxyPass wss://127.0.0.1:8001/main/realtime/sockjs/ + ProxyPassReverse wss://127.0.0.1:8001/main/realtime/sockjs/ + +EOF + + sudo -i <<'EOF' +export RENEWED_LINEAGE=/etc/letsencrypt/live/pump.iankelling.org +/a/bin/distro-setup/certbot-renew-hook +EOF + + s dd of=/etc/systemd/system/pump.service <<'EOF' +[Unit] +Description=pump.io +After=syslog.target network.target + +[Service] +Type=simple +User=pumpio +Group=pumpio +ExecStart=/home/iank/pump.io/bin/pump +Environment=NODE_ENV=production +# failed to find databank-mongodb without this. +# I just looked at my environment variables took a guess. +Environment=NODE_PATH=/usr/lib/nodejs:/usr/lib/node_modules:/usr/share/javascript + +[Install] +WantedBy=multi-user.target +EOF + ser daemon-reload + sgo pump + ########## end pump.io setup ############ + + + ############# begin setup mastodon ############## + + # main doc is Docker-Guide.md in docs repo + + # I'd like to try gnu social just cuz of gnu, but it's not being + # well maintained, for example, simple pull requests + # languishing: + # https://git.gnu.io/gnu/gnu-social/merge_requests/143 + # and I submitted my own bugs, basic docs are broken + # https://git.gnu.io/gnu/gnu-social/issues/269 + + # note, docker required, but we installed it earlier + + # i subscrubed to https://github.com/docker/compose/releases.atom + # to see release notes. + # i had some problems upgrading. blew things away with + # docker-compose down + # docker rmi $(docker images -q) + # s reboot now + # when running docker-compose run, kernel stack traces are printed to the journal. + # things seem to succeed, google says nothing, so ignoring them. + curl -L https://github.com/docker/compose/releases/download/1.18.0/docker-compose-`uname -s`-`uname -m` | s dd of=/usr/local/bin/docker-compose + s chmod +x /usr/local/bin/docker-compose + + + cd ~ + s rm -rf mastodon + i clone https://github.com/tootsuite/mastodon + cd mastodon + # subbed to atom feed to deal with updates + git checkout $(git tag | grep -v rc | tail -n1) + + # per instructions, uncomment redis/postgres persistence in docker-compose.yml + sed -i 's/^#//' docker-compose.yml + + cat >.env.production <<'EOF' +REDIS_HOST=redis +REDIS_PORT=6379 +DB_HOST=db +DB_USER=postgres +DB_NAME=postgres +DB_PASS= +DB_PORT=5432 + +LOCAL_DOMAIN=mast.iankelling.org +LOCAL_HTTPS=true + +SINGLE_USER_MODE=true + +SMTP_SERVER=mail.iankelling.org +SMTP_PORT=25 +SMTP_LOGIN=li +SMTP_FROM_ADDRESS=notifications@mast.iankelling.org +SMTP_DOMAIN=mast.iankelling.org +SMTP_DELIVERY_METHOD=smtp +EOF + + for key in PAPERCLIP_SECRET SECRET_KEY_BASE OTP_SECRET; do + # 1 minute 7 seconds to run this docker command + # to generate a secret, and it has ^M chars at the end. wtf. really dumb + printf "%s=%s\n" $key "$(docker-compose run --rm web rake secret|dos2unix|tail -n1)" >>.env.production + done + found=false + while read -r domain port pass; do + if [[ $domain == mail.iankelling.org ]]; then + found=true + # remove the username part + pass="${pass#*:}" + printf "SMTP_PASSWORD=%s\n" "$pass" >>.env.production + break + fi + done < <(s cat /etc/mailpass) + if ! $found; then + echo "$0: error, failed to find mailpass domain for mastadon" + exit 1 + fi + + # docker compose makes an interface named like br-8f3e208558f2. we need mail to + # get routed to us. + if ! s /sbin/iptables -t nat -C PREROUTING -i br-+ -p tcp -m tcp --dport 25 -j DNAT --to-destination 10.8.0.4:25; then + s /sbin/iptables -t nat -A PREROUTING -i br-+ -p tcp -m tcp --dport 25 -j DNAT --to-destination 10.8.0.4:25 + fi + + docker-compose run --rm web rake mastodon:webpush:generate_vapid_key | grep -E '^VAPID_PUBLIC_KEY=|^VAPID_PRIVATE_KEY=' >> .env.production + logq docker-compose run --rm web rake db:migrate + docker-compose run --rm web rails assets:precompile + + # avatar failed to upload, did + # docker logs mastodon_web_1 + # google lead me to this + s chown -R 991:991 public/system + + # docker daemon takes care of starting on boot. + docker-compose up -d + + s a2enmod proxy_wstunnel headers + web-conf -f 3000 - apache2 mast.iankelling.org <<'EOF' + ProxyPreserveHost On + RequestHeader set X-Forwarded-Proto "https" + ProxyPass /500.html ! + ProxyPass /oops.png ! + ProxyPass /api/v1/streaming/ ws://localhost:4000/ + ProxyPassReverse /api/v1/streaming/ ws://localhost:4000/ + ErrorDocument 500 /500.html + ErrorDocument 501 /500.html + ErrorDocument 502 /500.html + ErrorDocument 503 /500.html + ErrorDocument 504 /500.html +EOF + + + ############### !!!!!!!!!!!!!!!!! + ############### manual steps: + + # only following 2 people atm, so not bothering to figure out backups + # when mastodon has not documented it at all. + # + # fsf@status.fsf.org + # cwebber@toot.cat + # dbd@status.fsf.org + # johns@status.fsf.org + + # sign in page is at https://mast.iankelling.org/auth/sign_in + # register as iank, then + # https://github.com/tootsuite/documentation/blob/master/Running-Mastodon/Administration-guide.md + # docker-compose run --rm web bundle exec rails mastodon:make_admin USERNAME=iank + + ############# end setup mastodon ############## + + # we use nsupdate to update the ip of home + pi bind9 + + pi znc + # znc config generated by doing + # znc --makeconf + # selected port is also used in erc config + # comma separated channel list worked. + # while figuring things out, running znc -D for debug in foreground. + # to exit and save config: + # /msg *status shutdown + # configed auth on freenode by following + # https://wiki.znc.in/Sasl + # created the system service after, and had to do + # mv /home/iank/.znc/* /var/lib/znc + # sed -i 's,/home/iank/.znc/,/var/lib/znc,' /var/lib/znc/config/znc.conf + # and made a copy of the config files into /p/c + # added LoadModule = log -sanitize to the top level + # to get into the web interface, + # cat /etc/letsencrypt/live/iankelling.org/{privkey,cert,chain}.pem > /var/lib/znc/znc.pem + # then use non-main browser or else it doesn't allow it based on ocsp stapling from my main site. + # i'm going to figure out how to automate this when it expires. i know i can hook a script into the renewal. https://wiki.znc.in/FAQ seems to imply that znc doesn\'t need restart. + # todo: in config file AllowWeb = true should be false. better security if that is off unless we need it. + # todo: figure out how to make playback in erc happe.n + s useradd --create-home -d /var/lib/znc --system --shell /sbin/nologin --comment "Account to run ZNC daemon" --user-group znc || [[ $? == 9 ]] # 9 if it exists already + chmod 700 /var/lib/znc + s chown -R znc:znc /var/lib/znc/config + s dd of=/etc/systemd/system/znc.service 2>/dev/null <<'EOF' +[Unit] +Description=ZNC, an advanced IRC bouncer +After=network-online.target + +[Service] +ExecStart=/usr/bin/znc -f --datadir=/var/lib/znc +User=znc + +[Install] +WantedBy=multi-user.target +EOF + ser daemon-reload + sgo znc + + echo "$0: $(date): ending now)" + exit 0 + ;; esac +# needed for checkrestart +if isdeb; then + spa debian-goodies +fi + + + +########### end section including li/lj ############### + +case $distro in + debian) spa gnome-session-flashback ;; + # flidas is missing dependency gnome-panel. others unknown +esac + + + +case $distro in + trisquel|ubuntu|debian) spa ack-grep ;; + arch|fedora) spa ack ;; + # fedora unknown +esac + + + +case $distro in + debian) + pi chromium ;; + xenial|ubuntu) + wget -qO - https://downloads.iridiumbrowser.de/ubuntu/iridium-release-sign-01.pub|sudo apt-key add - + cat < false, +'rpc-authentication-required' => false, +'incomplete-dir' => '/i/k/partial-torrents', +'incomplete-dir-enabled' => true, +'download-dir' => '/i/k/torrents', +"speed-limit-up" => 800, +"speed-limit-up-enabled" => true, +"peer-port" => 61486, +"cache-size-mb" => 256, +"ratio-limit" => 5.0, +"ratio-limit-enabled" => true, })) + "\n") EOF - sgo transmission-daemon - ;; - arch) - # todo, setup it's config file & daemon - pi transmission-cli - ;; - esac -fi + # make sure its not enabled, not sure if this is needed + ser disable transmission-daemon + ;; + # todo: others unknown +esac # adapted from /var/lib/dpkg/info/transmission-daemon.postinst if ! getent passwd debian-transmission > /dev/null; then case $distro in @@ -188,284 +1157,193 @@ if ! getent passwd debian-transmission > /dev/null; then esac fi -# arch had a default config, -# debian had nothing until you start it. -# With a little trial an error, here is a minimal config -# taken from the generated one, plus changes that the -# settings ui does, without a bunch of ui crap settings. -# -# only settings I set were -# hostname -# auto-connect -# -for f in /home/*; do - d=$f/.config/transmission-remote-gtk - u=${f##*/} - s -u $u mkdir -p $d - s -u $u dd of=$d/config.json <<'EOF' -{ - "profiles" : [ - { - "profile-name" : "Default", - "hostname" : "frodo", - "rpc-url-path" : "/transmission/rpc", - "username" : "", - "password" : "", - "auto-connect" : true, - "ssl" : false, - "timeout" : 40, - "retries" : 3, - "update-active-only" : false, - "activeonly-fullsync-enabled" : false, - "activeonly-fullsync-every" : 2, - "update-interval" : 3, - "min-update-interval" : 3, - "session-update-interval" : 60, - "exec-commands" : [ - ], - "destinations" : [ - ] - } - ], - "profile-id" : 0, - "add-options-dialog" : false -} -EOF -done -case $distro in - debian|ubuntu) - pi-nostart openvpn - # pi-nostart this doesn't seem to be good enough? - ser disable openvpn@client - ser disable openvpn - ;; - *) pi openvpn ;; - esac - - case $HOSTNAME in - tp|frodo) - case $distro in - debian|ubuntu) - log=$(mktemp) - cd /a/opt - wget -N https://dl.google.com/linux/direct/google-chrome-stable_current_amd64.deb - set +e - s dpkg -i google-chrome-stable_current_amd64.deb &> $log - code=$? - set -e - case $code in - 1) - if grep '^dpkg: dependency problems prevent configuration of' \ - $log &>/dev/null; then - s apt-get -fy install - else - exit 1 - fi - ;; - 0) : ;; - *) exit $code - esac - ;; - arch) - pi google-chrome - ;; - esac - ;; - esac +# trisquel 8 = openvpn, debian stretch = openvpn-client +vpn_ser=openvpn-client +if [[ ! -e /lib/systemd/system/openvpn-client@.service ]]; then + vpn_ser=openvpn +fi - case $distro in - # ubuntu unknown. probably the same as debian, just check if the - # init scripts come with the package. - debian) - # copied from arch, but moved to etc - s dd of=/etc/systemd/user/synergys.service <<'EOF' +s dd of=/etc/systemd/system/transmission-daemon-nn.service < /tmp/x -fi - +# btrfs maintenance +sgo btrfsmaint.timer +sgo btrfsmaintstop.timer -######### begin postfix ######## -# based on,http://www.postfix.org/qmgr.8.html and my notes in gnus -# originally tried moving specific directories under /var/spool/postfix, -# but postfix didn't like that -if [[ ! -L /var/spool/postfix ]]; then - ser stop postfix - n=/q/postfix-`distro-name``debian-archive` - if [[ -e $n ]]; then - echo "$0: warning: $n already exists before we do the link, removing it" - rm -rf $n - fi - s mv /var/spool/postfix $n - s lnf -T $n /var/spool/postfix - ser start postfix - journalctl -n 20 | cat -fi +# the wiki backup script from ofswiki.org uses generic paths +s lnf /p/c/machine_specific/li/mw_vars /root +s lnf /k/backup/wiki_backup /root -# This also works instead of ~/.forward -# s sed -i '/^root/d' /etc/aliases ||: -#echo "root: $HOSTNAME@bog.mm.st" | s tee -a /etc/aliases -# this can't be a symlink and has permission restrictions -# it might work in /etc/aliases, but this seems more proper. +s cedit /etc/goaccess.conf <<'EOF' || [[ $? == 1 ]] +# all things found from looking around the default config +# copied existing NCSA Combined Log Format with Virtual Host, plus %L +log-format %^:%^ %h %^[%d:%t %^] "%r" %s %b "%R" "%u" %D +time-format %H:%M:%S +date-format %d/%b/%Y +log-file /var/log/apache2/access.log +color-scheme 2 -if s grep amazonaws /etc/postfix/sasl_passwd &>/dev/null; then - forward=$HOSTNAME@sallymae.club -else - forward=$HOSTNAME@bog.mm.st -fi -e $forward > ~/.forward -e $forward | s tee /root/.forward -s newaliases - -# if I wanted the from address to be renamed and sent to a different address, -# echo "sdx@localhost development@localhost" | sudo dd of=/etc/postfix/recipient_canonical -# sudo postmap hash:/etc/postfix/recipient_canonical -# sudo service postfix reload - - -# i'm assuming mail just won't work on systems without the sasl_passwd. -postconfin <<'EOF' -smtp_sasl_auth_enable = yes -smtp_sasl_password_maps = hash:/etc/postfix/sasl_passwd -smtp_sasl_security_options = noanonymous -smtp_tls_security_level = secure -message_size_limit = 20480000 -smtp_tls_CAfile = /etc/ssl/certs/ca-certificates.crt +# tip: copy access.log files to a stretch host directory, then run +# jessie's goaccess is too old for some options, and it\'s +# not easily installed from a testing. +# goaccess --ignore-crawlers -f <(cat *) -a -o html > x.html EOF -# ^ I ran into a log file not sending cuz of size. double from 10 to 20 meg limit -s postmap hash:/etc/postfix/sasl_passwd -# offlineimap uses this too, it is much easier to use one location than to -# condition it's config and postfix's config + +case $distro in + trisquel|ubuntu|debian) + # unison-gtk second, i want it to be default, not sure if that works + # with spa. note, I used to install from testing repo when using stable, + # but it shouldn't be needed since I wrote a script to handle mismatching + # compilers. + spa unison unison-gtk + ;; +arch) + spa unison gtk2 + ;; +esac + case $distro in - fedora) s lnf -T ca-certificates.crt /etc/ssl/ca-bundle.trust.crt ;; - *) : + arch) + # default is alsa, doesn\'t work with with pianobar + s dd of=/etc/libao.conf <<'EOF' +default_driver=pulse +EOF + ;; esac -s service postfix reload -sgo postfix +# note, for jessie, it depends on a higher version of btrfs-tools. +# +# # disabled due to my patch being in btrbk +# case $distro in +# arch|debian|trisquel|ubuntu) pi btrbk ;; +# # others unknown +# esac +cd /a/opt/btrbk +s make install +spa pv # for progress bar when running interactively. + +# ian: temporarily disabled while hosts are in flux. +# if [[ $HOSTNAME == tp ]]; then +# # backup/sync manually on others hosts for now. +# sgo btrbk.timer +# # note: to see when it was last run, +# # ser list-timers +# fi + -############ end postfix ####### case $distro in - debian|ubuntu) s gpasswd -a ian adm ;; #needed for reading logs + debian|trisquel|ubuntu) s gpasswd -a iank adm ;; #needed for reading logs esac # tor @@ -900,7 +1809,7 @@ EOF pi nfs-utils sgo nfs-server ;; - debian|ubuntu) + debian|trisquel|ubuntu) pi nfs-server ;; arch) @@ -911,150 +1820,57 @@ EOF ;; esac -if [[ -e /i/video ]]; then + + + +if [[ $HOSTNAME == frodo ]]; then # nohide = export filesystems mounted deeper than the export point # fsid=0 makes this export the "root" export # not documented in the man page, but this means # 1. it can be mounted with a shorthand of server:/ # 2. exports that are subdirectories of this one will automatically be mounted - tu /etc/exports '/i/video 192.168.1.0/24(rw,fsid=0,nohide,no_root_squash,async,no_subtree_check,insecure)' + tu /etc/exports <<'EOF' +/k 192.168.1.0/24(rw,fsid=0,nohide,no_root_squash,async,no_subtree_check,insecure) +EOF s exportfs -rav - showmount -e localhost fi -# cron -f=/a/bin/$HOSTNAME-crontab -if [[ -e $f ]]; then - $f -fi - e "$end_msg_var" # persistent virtual machines - case $distro in - debian|ubuntu) + debian|trisquel|ubuntu) pi libosinfo-bin; ;; esac +# if I was going to create a persistent vm, i might do it like this: +# variant=something # from: virt-install --os-variant list +# s virt-install --noautoconsole --graphics spice,listen=0.0.0.0 \ + # --disk=/a/images/some_name.qcow2,bus=virtio --vcpus 2 -r 4096 -w bridge=br0 \ + # -n some_name --import --os-variant $variant --cpu host-model-only -# distro may not know about win 10 yet. -variant=win7 -if ! virt-install --os-variant list &>/dev/null; then # we are using a newer virt-install - for v in 10 8.1 8; do - if osinfo-query os | gr "^\s*win${v/./\\.}\s" &>/dev/null; then - variant=win$v - break - fi - done -fi - -if ! s virsh list --all --name | grep -xF win10 &>/dev/null; then - - # created account with - # win10vmian@outlook.com, and easy to remember password - # win 10 virtio, makes disk way way way faster - # wget https://fedorapeople.org/groups/virt/virtio-win/direct-downloads/latest-virtio/virtio-win.iso - # https://wiki.archlinux.org/index.php/QEMU#Change_Existing_Windows_VM_to_use_virtio - # for installing virtio after initial install instead of with initial iso: - # qemu-img create -f qcow2 fake.qcow2 1G - # --disk=/a/images/virtio-win.iso,device=cdrom \ - # --disk=/a/images/fake.qcow2,bus=virtio - # Also, - # went to device manager, saw 2 pci devices with yellow !, - # did search for drivers, pick cdrom location, done. - # - # from http://www.tenforums.com/tutorials/4189-fast-startup-turn-off-windows-10-a.html. - # google said there was a control panel option for it, but - # that turned out to be a lie. - # Put this in a .bat file and run as administrator to turn off - # hyberboot which fucks things up. - # REG ADD "HKLM\SYSTEM\CurrentControlSet\Control\Session Manager\Power" /V HiberbootEnabled /T REG_dWORD /D 0 /F - # power settings, turn off display: never - # run "control userpasswords2", turn on automatic login. - # note: when changing devices, I just undefine, the create the vm again. - - s virt-install --noautoconsole --graphics spice,listen=0.0.0.0 \ - --disk=/a/images/win10.qcow2,bus=virtio --vcpus 2 -r 4096 -w bridge=br0 \ - -n win10 --import --os-variant $variant --cpu host-model-only - - s virsh destroy win10 - - # this one hasn\'t had the virtio fix done yet. - s virt-install --noautoconsole --graphics spice,listen=0.0.0.0 \ - --disk=/a/images/win7.qcow2 --vcpus 2 -r 4096 -w bridge=br0 \ - -n win7 --import --os-variant win7 --cpu host-model-only - s virsh destroy win7 - # had a problem with --cpu host, so trying out - # --cpu host-model-only - -fi -pi samba -# note samba re-reads it's config every 1 minute -case $distro in - arch) s cp /etc/samba/smb.conf.default /etc/samba/smb.conf ;; -esac +######### begin stuff belonging at the end ########## -# add 2 lines after workgroup option -s sed -ri '/^\s*encrypt passwords\s*=/d' /etc/samba/smb.conf -s sed -ri '/^\s*map to guest\s*=/d' /etc/samba/smb.conf -s sed -i 's/\(\s*workgroup\s*=\).*/\1 WORKGROUP\n\tencrypt passwords = yes\n\tmap to guest = bad password/' /etc/samba/smb.conf -# remove default homes section. not sharing that. -s sed -ri '/^\s*\[homes\]/,/\s*\[/d' /etc/samba/smb.conf - -if ! grep -xF '[public]' /etc/samba/smb.conf &>/dev/null; then - s tee -a /etc/samba/smb.conf <<'EOF' -[public] - guest ok = yes - read only = no - path = /kfrodo -EOF -fi case $distro in - debian|ubuntu) -# systemd claims it generates units from /etc/init.d, but it clearly doesn't -# in debian. I have no idea how they are related. fuck debian right now. It's -# not documented. samba has a systemd init file linked to /dev/null. -# There's this https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=769714 which -# claims samba's sub-services will be started automatically by systemd... it -# didn't on install, wonder if it will on boot. It clued me in how to start -# it manually though. Nothing in /usr/share/doc/samba, debian admin guide -# says nothing about any of this. (this is in debian testing as of 4/2016). - -s /etc/init.d/samba start -;; + ubuntu|debian) + spa spacefm-gtk3 ;; arch) - sgo samba - ;; + spa spacefm ;; esac -tu /etc/hosts <<< "127.0.1.1 $(hostname).lan $(hostname)" +pi "${simple_packages[@]}" -rootdev=$(mount | sed -rn 's#^(\S+) on / .*#\1#p') -s mkdir /mnt/root -tu /etc/fstab <<< "$rootdev /mnt/root btrfs noatime,subvolid=0 0 0" -mountpoint /mnt/root || s mount /mnt/root -idev=$(mount | sed -rn 's#^(\S+) on /i .*#\1#p') -if [[ $idev != $rootdev ]]; then - s mkdir /mnt/iroot - tu /etc/fstab <<< "$idev /mnt/iroot btrfs noatime,subvolid=0 0 0" - mountpoint /mnt/iroot || s mount /mnt/iroot -fi - -# Do this again because it occasionally has changes and -# it can be run outside initial isntall. -s /a/bin/fai/fai/config/distro-install-common/end - if $pending_reboot; then echo "$0: pending reboot and then finished. doing it now." s reboot now else echo "$0: $(date): ending now)" fi +exit 0