]> git.openstreetmap.org Git - chef.git/blobdiff - cookbooks/planet/templates/default/planetdump.erb
Stop handling X-Forwarded-For from old tile CDN
[chef.git] / cookbooks / planet / templates / default / planetdump.erb
index 9df4a6f0372c8222dd1a9dda870e0c38e1a41350..f0708f2f26fa3fc625273cb59c06b607b7dc7b32 100644 (file)
 #!/bin/bash
 #!/bin/bash
-ulimit -m 4194304 -v 4194304
-set -e
-
-if [ -f /tmp/planetdump.lock ]; then
-       if [ "$(ps -p `cat /tmp/planetdump.lock` | wc -l)" -gt 1 ]; then
-               # process is still running
-               echo Error: Another planetdump is running
-               exit 1
-       else
-               # process not running, but lock file not deleted?
-               rm /tmp/planetdump.lock
-       fi
-fi
-
-#Create Lock
-echo $$ > /tmp/planetdump.lock
 
 
+# DO NOT EDIT - This file is being maintained by Chef
 
 
-cur_date=`date +%y%m%d`
-cur_year=`date +%Y`
-cur_planet=planet-${cur_date}.osm
-cur_changeset=changesets-${cur_date}.osm
-planet_dir=/store/planet/
-
-export CONNECTION_PARAMS='dbname=openstreetmap host=db user=planetdump password=<%= @password %>'
-export PATH='/usr/local/bin:/usr/bin:/bin:/usr/bin/X11'
+# Exit on error
+set -e
 
 
+# Get the name of the file and the expected pattern
+file="$1"
+pattern="^osm-([0-9]{4})-([0-9]{2})-([0-9]{2})\.dmp$"
 
 
-if [ \! -d ${planet_dir}/planet/ ]
-       then
-               echo ${planet_dir}planet/ does not exist
-               exit 1
-       fi
-if [ \! -d ${planet_dir}/planet/${cur_year}/ ]; then mkdir ${planet_dir}/planet/${cur_year}/; fi
-cd ${planet_dir}/planet/${cur_year}/
+# Give up now if the file isn't a database dump
+[[ $file =~ $pattern ]] || exit 0
 
 
-/opt/planetdump/planet06_pg | pbzip2 -p6 -6c > .${cur_planet}.bz2.new
-/opt/planetdump/planet06_pg --changesets | pbzip2 -p6 -9c > .${cur_changeset}.bz2.new
+# Save the year and date from the file name
+year="${BASH_REMATCH[1]}"
+date="${year:2:2}${BASH_REMATCH[2]}${BASH_REMATCH[3]}"
 
 
-planet_size=$(du -sb .${cur_planet}.bz2.new | awk '{ print $1 }')
-changeset_size=$(du -sb .${cur_changeset}.bz2.new | awk '{ print $1 }')
-if ((planet_size<32000000000)); then
-       echo Planet .${cur_planet}.bz2.new too small
-       exit 1
-fi
-if ((changeset_size<720000000)); then
-       echo Changeset .${cur_changeset}.bz2.new too small
-       exit 1
+# Check the lock
+if [ -f /tmp/planetdump.lock ]; then
+    if [ "$(ps -p `cat /tmp/planetdump.lock` | wc -l)" -gt 1 ]; then
+        echo "Error: Another planetdump is running"
+        exit 1
+    else
+        rm /tmp/planetdump.lock
+    fi
 fi
 
 fi
 
-mv .${cur_planet}.bz2.new ${cur_planet}.bz2
-mv .${cur_changeset}.bz2.new ${cur_changeset}.bz2
-md5sum ${cur_planet}.bz2 > ${cur_planet}.bz2.md5
-md5sum ${cur_changeset}.bz2 > ${cur_changeset}.bz2.md5
-
-#link planet latest to the new file
-cd ${planet_dir}/planet/
-
-ln -fs ${cur_year}/${cur_planet}.bz2 planet-latest.osm.bz2
-ln -fs ${cur_year}/${cur_changeset}.bz2 changesets-latest.osm.bz2
+# Redirect this shell's output to a file. This is so that it
+# can be emailed later, since this script is run from incron
+# and incron doesn't yet support MAILTO like cron does. The
+# command below appears to work in bash as well as dash.
+logfile="/tmp/planetdump.log.$$"
+exec > "${logfile}" 2>&1
 
 
-# mangle md5 files for 'latest' ones
-rm -f planet-latest.osm.bz2.md5
-rm -f changesets-latest.osm.bz2.md5
-
-sed -e "s/${cur_planet}.bz2/planet-latest.osm.bz2/" ${cur_year}/${cur_planet}.bz2.md5 > planet-latest.osm.bz2.md5
-sed -e "s/${cur_changeset}.bz2/changesets-latest.osm.bz2/" ${cur_year}/${cur_changeset}.bz2.md5 > changesets-latest.osm.bz2.md5
-
-rm /tmp/planetdump.lock
+# Create lock file
+echo $$ > /tmp/planetdump.lock
 
 
+# Define cleanup function
+function cleanup {
+    # Remove the lock file
+    rm /tmp/planetdump.lock
+
+    # Send an email with the output, since incron doesn't yet
+    # support doing this in the incrontab
+    if [[ -s "$logfile" ]]
+    then
+        mailx -s "Planet dump output: ${file}" zerebubuth@gmail.com < "${logfile}"
+    fi
+
+    # Remove the log file
+    rm -f "${logfile}"
+}
+
+# Remove lock on exit
+trap cleanup EXIT
+
+# Change to working directory
+cd /store/planetdump
+
+# Cleanup
+rm -rf users
+rm -rf changesets changeset_tags changeset_comments
+rm -rf nodes node_tags
+rm -rf ways way_tags way_nodes
+rm -rf relations relation_tags relation_members
+
+# Run the dump
+time nice -n 19 /opt/planet-dump-ng/planet-dump-ng \
+     --max-concurrency=4 \
+     -c "pbzip2 -c" -f "/store/backup/${file}" --dense-nodes=1 \
+     -C "changesets-${date}.osm.bz2" \
+     -D "discussions-${date}.osm.bz2" \
+     -x "planet-${date}.osm.bz2" -X "history-${date}.osm.bz2" \
+     -p "planet-${date}.osm.pbf" -P "history-${date}.osm.pbf"
+
+# Function to create bittorrent files
+function mk_torrent {
+  type="$1"
+  format="$2"
+  dir="$3"
+  s_year="$4"
+  web_dir="${dir}${s_year}"
+  name="${type}-${date}.osm.${format}"
+  web_path="${web_dir}/${name}"
+  rss_web_dir="https://planet.openstreetmap.org/${dir}"
+  rss_file="${type}-${format}-rss.xml"
+  torrent_file="${name}.torrent"
+  torrent_url="${rss_web_dir}${s_year}/${torrent_file}"
+
+  # create .torrent file
+  mktorrent -l 22 "${name}" \
+     -a udp://tracker.opentrackr.org:1337 \
+     -a udp://tracker.datacenterlight.ch:6969/announce,http://tracker.datacenterlight.ch:6969/announce \
+     -a udp://tracker.torrent.eu.org:451 \
+     -a udp://tracker-udp.gbitt.info:80/announce,http://tracker.gbitt.info/announce,https://tracker.gbitt.info/announce \
+     -a http://retracker.local/announce \
+     -w "https://planet.openstreetmap.org/${web_path}" \
+     -w "https://ftp5.gwdg.de/pub/misc/openstreetmap/planet.openstreetmap.org/${web_path}" \
+     -w "https://ftpmirror.your.org/pub/openstreetmap/${web_path}" \
+     -w "https://mirror.init7.net/openstreetmap/${web_path}" \
+     -w "https://free.nchc.org.tw/osm.planet/${web_path}" \
+     -w "https://ftp.fau.de/osm-planet/${web_path}" \
+     -w "https://ftp.spline.de/pub/openstreetmap/${web_path}" \
+     -w "https://osm.openarchive.site/${name}" \
+     -w "https://downloads.opencagedata.com/planet/${name}" \
+     -w "https://planet.osm-hr.org/${web_path}" \
+     -w "https://planet.maps.mail.ru/${web_path}" \
+     -c "OpenStreetMap ${type} data export, licensed under https://opendatacommons.org/licenses/odbl/ by OpenStreetMap contributors" \
+     -o "${torrent_file}" > /dev/null
+
+  # create .xml global RSS headers if missing
+  torrent_time_rfc="$(date -R -r ${torrent_file})"
+  test -f "${rss_file}" || echo "<x/>" | xmlstarlet select --xml-decl --indent \
+       -N "atom=http://www.w3.org/2005/Atom" \
+       -N "dcterms=http://purl.org/dc/terms/" \
+       -N "content=http://purl.org/rss/1.0/modules/content/" \
+       --encode "UTF-8" \
+       --template \
+       --match / \
+       --elem "rss" \
+               --attr "version" --output "2.0" --break \
+               --attr "atom:DUMMY" --break \
+       --elem "channel" \
+       --elem "title" --output "OpenStreetMap ${type} ${format} torrent RSS" --break \
+       --elem "link"  --output "${rss_web_dir}" --break \
+       --elem "atom:link" \
+               --attr "href" --output "${rss_web_dir}/${rss_file}" --break \
+               --attr "rel" --output "self" --break \
+               --attr "type" --output "application/rss+xml" --break \
+               --break \
+       --elem "description" --output "${type}.osm.${format}.torrent RSS feed" --break \
+       --elem "copyright" --output "Source: OpenStreetMap contributors, under ODbL 1.0 licence" --break \
+       --elem "generator" --output "OpenStreetMap xmlstarlet powered shell script v1.0" --break \
+       --elem "language" --output "en" --break \
+       --elem "lastBuildDate" --output "${torrent_time_rfc}" \
+       > "${rss_file}"
+
+  # add newly created .torrent file as new entry to .xml RSS feed, removing excess entries
+  torrent_size="$(stat --format="%s" ${torrent_file})"
+  xmlstarlet edit --inplace \
+       -a "//lastBuildDate" -t elem -n item -v ""  \
+       -s "//item[1]" -t elem -n "title" -v "${torrent_file}" \
+       -s "//item[1]" -t elem -n "guid" -v "${torrent_url}" \
+       -s "//item[1]" -t elem -n "link" -v "${torrent_url}" \
+       -s "//item[1]" -t elem -n "pubDate" -v "${torrent_time_rfc}" \
+       -s "//item[1]" -t elem -n "category" -v "OpenStreetMap data" \
+       -s "//item[1]" -t elem -n "enclosure" \
+               -s "//item[1]"/enclosure -t attr -n "type" -v "application/x-bittorrent" \
+               -s "//item[1]"/enclosure -t attr -n "length" -v "${torrent_size}" \
+               -s "//item[1]"/enclosure -t attr -n "url" -v "${torrent_url}" \
+       -s "//item[1]" -t elem -n "description" -v "OpenStreetMap torrent ${torrent_file}" \
+       -u /rss/channel/lastBuildDate -v "${torrent_time_rfc}" \
+       -d /rss/@atom:DUMMY \
+       -d "//item[position()>5]" \
+       "${rss_file}"
+}
+
+# Function to install a dump in place
+function install_dump {
+  type="$1"
+  format="$2"
+  dir="$3"
+  year="$4"
+  name="${type}-${date}.osm.${format}"
+  latest="${type}-latest.osm.${format}"
+  rss_file="${type}-${format}-rss.xml"
+
+  md5sum "${name}" > "${name}.md5"
+  mkdir -p "${dir}/${year}"
+  mv "${name}" "${name}.md5" "${dir}/${year}"
+  ln -sf "${year:-.}/${name}" "${dir}/${latest}"
+  test -f "${name}.torrent" && mv "${name}.torrent" "${dir}/${year}" && ln -sf "${year:-.}/${name}.torrent" "${dir}/${latest}.torrent"
+  test -f "${rss_file}" && xmllint --noout "${rss_file}" && cp -f "${rss_file}" "${dir}"
+  rm -f "${dir}/${latest}.md5"
+  sed -e "s/${name}/${latest}/" "${dir}/${year}/${name}.md5" > "${dir}/${latest}.md5"
+}
+
+# Create *.torrent files
+mk_torrent "changesets" "bz2" "planet" "/${year}"
+mk_torrent "discussions" "bz2" "planet" "/${year}"
+mk_torrent "planet" "bz2" "planet" "/${year}"
+mk_torrent "history" "bz2" "planet/full-history" "/${year}"
+mk_torrent "planet" "pbf" "pbf"
+mk_torrent "history" "pbf" "pbf/full-history"
+
+# Move dumps into place
+install_dump "changesets" "bz2" "<%= node[:planet][:dump][:xml_directory] %>" "${year}"
+install_dump "discussions" "bz2" "<%= node[:planet][:dump][:xml_directory] %>" "${year}"
+install_dump "planet" "bz2" "<%= node[:planet][:dump][:xml_directory] %>" "${year}"
+install_dump "history" "bz2" "<%= node[:planet][:dump][:xml_history_directory] %>" "${year}"
+install_dump "planet" "pbf" "<%= node[:planet][:dump][:pbf_directory] %>"
+install_dump "history" "pbf" "<%= node[:planet][:dump][:pbf_history_directory] %>"
+
+# Remove pbf dumps older than 90 days
+find "<%= node[:planet][:dump][:pbf_directory] %>" "<%= node[:planet][:dump][:pbf_history_directory] %>" \
+     -maxdepth 1 -mindepth 1 -type f -mtime +90 \
+     \( \
+     -iname 'planet-*.pbf' \
+     -o -iname 'history-*.pbf' \
+     -o -iname 'planet-*.pbf.md5' \
+     -o -iname 'history-*.pbf.md5' \
+     \) \
+     -delete