X-Git-Url: https://git.openstreetmap.org./chef.git/blobdiff_plain/8771d94d8bfd599356ab8a6f697289a26788a9b3..d18cd92a98384aebf371376383a4408485f12374:/cookbooks/planet/templates/default/planetdump.erb

diff --git a/cookbooks/planet/templates/default/planetdump.erb b/cookbooks/planet/templates/default/planetdump.erb
index d60de2b5b..5c0d38045 100644
--- a/cookbooks/planet/templates/default/planetdump.erb
+++ b/cookbooks/planet/templates/default/planetdump.erb
@@ -1,71 +1,204 @@
 #!/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-slave 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
 
-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
-
-# 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
+}
+
+# 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 {
+  local type="$1"
+  local format="$2"
+  local dir="$3"
+  local s3path="$4"
+  local s_year="$5"
+  local web_dir="${dir}${s_year}"
+  local name="${type}-${date}.osm.${format}"
+  local web_path="${web_dir}/${name}"
+  local s3_web_path="${s3path}/${name}"
+  local rss_web_dir="https://planet.openstreetmap.org/${dir}"
+  local rss_file="${type}-${format}-rss.xml"
+  local torrent_file="${name}.torrent"
+  local 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/${web_path}" \
+     -w "https://osm-planet-eu-central-1.s3.dualstack.eu-central-1.amazonaws.com/${s3_web_path}" \
+     -w "https://osm-planet-us-west-2.s3.dualstack.us-west-2.amazonaws.com/${s3_web_path}" \
+     -w "https://ftp5.gwdg.de/pub/misc/openstreetmap/planet.openstreetmap.org/${web_path}" \
+     -w "https://ftpmirror.your.org/pub/openstreetmap/${web_path}" \
+     -w "https://mirror.init7.net/openstreetmap/${web_path}" \
+     -w "https://ftp.fau.de/osm-planet/${web_path}" \
+     -w "https://ftp.spline.de/pub/openstreetmap/${web_path}" \
+     -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 replication_status_wait {
+  local s3_url="$1"
+  for i in {1..3600}; do
+    local replication_status=$(curl -sI --location "${s3_url}" | grep -F 'x-amz-replication-status' | awk '{print $2}' |  tr -d '\r' )
+
+    if [[ "${replication_status}" == "COMPLETED" ]]; then
+      return 0 # success
+    fi
+
+    sleep 1
+  done
+  echo "Timeout waiting for ${s3_url} to complete replication status: ${replication_status}"
+}
+
+# Function to install a dump in place
+function install_dump {
+  local type="$1"
+  local format="$2"
+  local dir="$3"
+  local s3dir="$4"
+  local year="$5"
+  local name="${type}-${date}.osm.${format}"
+  local latest="${type}-latest.osm.${format}"
+  local rss_file="${type}-${format}-rss.xml"
+
+  md5sum "${name}" > "${name}.md5"
+
+  # Upload all files to S3
+  /opt/awscli/v2/current/bin/aws --profile=osm-pds-upload s3 cp --storage-class=INTELLIGENT_TIERING --no-progress "${name}.md5" "s3://osm-planet-eu-central-1/${s3dir}/${name}.md5"
+  /opt/awscli/v2/current/bin/aws --profile=osm-pds-upload s3 cp --storage-class=INTELLIGENT_TIERING --no-progress "${name}.torrent" "s3://osm-planet-eu-central-1/${s3dir}/${name}.torrent"
+  /opt/awscli/v2/current/bin/aws --profile=osm-pds-upload s3 cp --storage-class=INTELLIGENT_TIERING --no-progress "${name}" "s3://osm-planet-eu-central-1/${s3dir}/${name}"
+
+  # Waiting for S3 replication to complete
+  replication_status_wait "https://osm-planet-eu-central-1.s3.dualstack.eu-central-1.amazonaws.com/${s3dir}/${name}.md5"
+  replication_status_wait "https://osm-planet-eu-central-1.s3.dualstack.eu-central-1.amazonaws.com/${s3dir}/${name}.torrent"
+  replication_status_wait "https://osm-planet-eu-central-1.s3.dualstack.eu-central-1.amazonaws.com/${s3dir}/${name}"
+
+  mkdir -p "${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" "changesets/osm/${year}" "/${year}"
+mk_torrent "discussions" "bz2" "planet" "discussions/osm/${year}" "/${year}"
+mk_torrent "planet" "bz2" "planet" "planet/osm/${year}" "/${year}"
+mk_torrent "history" "bz2" "planet/full-history" "planet-full-history/osm/${year}" "/${year}"
+mk_torrent "planet" "pbf" "pbf" "planet/pbf/${year}"
+mk_torrent "history" "pbf" "pbf/full-history" "planet-full-history/pbf/${year}"
+
+# Move dumps into place
+install_dump "changesets" "bz2" "<%= node[:planet][:dump][:xml_directory] %>" "changesets/osm/${year}" "${year}"
+install_dump "discussions" "bz2" "<%= node[:planet][:dump][:xml_directory] %>" "discussions/osm/${year}" "${year}"
+install_dump "planet" "bz2" "<%= node[:planet][:dump][:xml_directory] %>" "planet/osm/${year}" "${year}"
+install_dump "history" "bz2" "<%= node[:planet][:dump][:xml_history_directory] %>" "planet-full-history/osm/${year}" "${year}"
+install_dump "planet" "pbf" "<%= node[:planet][:dump][:pbf_directory] %>" "planet/pbf/${year}"
+install_dump "history" "pbf" "<%= node[:planet][:dump][:pbf_history_directory] %>" "planet-full-history/pbf/${year}"