Add from Parabola: db-cleanup

If you're seeing this in `git blame`, it chose to follow the wrong
ancestor of a merge commit.

5 files changed, 91 insertions, 0 deletions

diff --git a/README.md b/README.md
index 4072a6a..7522376 100644
--- a/README.md
+++ b/README.md
@@ -8,6 +8,7 @@ The executables that you (might) care about are:
 
     dbscripts/
     ├── cron-jobs/
+    │   ├── db-cleanup                  [Parabola only]
     │   ├── devlist-mailer
     │   ├── ftpdir-cleanup
     │   ├── integrity-check
@@ -44,6 +45,11 @@ When we remove a package from a repository, it stays in the package
 the pool, to reclaim the disk space:
 
  - `cron-jobs/ftpdir-cleanup`
+ - `cron-jobs/db-cleanup`
+
+Both of these programs do the exact same thing.  Parabola developers
+decided to write their own from scratch, instead of modifying
+`ftpdir-cleanup`.  They should eventually be merged.
 
 Things that haven't been mentioned yet:
 
diff --git a/cron-jobs/db-cleanup b/cron-jobs/db-cleanup
new file mode 100755
index 0000000..04a559a
--- /dev/null
+++ b/cron-jobs/db-cleanup
@@ -0,0 +1,48 @@
+#!/bin/bash
+# Syncs pools against themselves using database contents as filter to cleanup
+# them up
+# License: GPLv3
+
+# Principles
+# * Get repos dbs contents
+# * Make them a include list
+# * Rsync pools against themselves removing excluded files
+# * Instant cleanup!
+
+set -eu -o pipefail
+source "$(dirname "$(readlink -e "$0")")/../config"
+source "$(dirname "$(readlink -e "$0")")/../db-cleanup.conf"
+source "$(librelib messages)"
+setup_traps
+
+EXTRAFLAGS=()
+if [[ $CLEANUP_DRYRUN = true ]]; then
+	EXTRAFLAGS+=(--dry-run)
+fi
+
+filter=$(mktemp -t "${0##*/}.XXXXXXXXXX")
+trap "rm -f -- ${filter@Q}" EXIT
+
+for dbfile in "${FTP_BASE}"/*/os/*/*.db; do
+	msg 'Processing %s' "$dbfile"
+	bsdtar tf "${dbfile}"
+done | cut -d'/' -f1 | sort -u | sed "s|$|*|" > "$filter"
+
+msg "Removing old files:"
+
+for POOL in "${PKGPOOLS[@]}" "${SRCPOOLS[@]}"; do
+	msg2 '%s' "${POOL}"
+
+	rsync "${EXTRAFLAGS[@]}" -va --delete-excluded \
+		--include-from="$filter" \
+		--exclude="*" \
+		"${FTP_BASE}/${POOL}/" \
+		"${FTP_BASE}/${POOL}/"
+done
+
+msg "Removing dead symlinks:"
+actions=(-print)
+if [[ $CLEANUP_DRYRUN != true ]]; then
+	actions+=(-delete)
+fi
+find -L "${FTP_BASE}/" -type l "${actions[@]}"
diff --git a/db-cleanup.conf b/db-cleanup.conf
new file mode 100644
index 0000000..12d65f9
--- /dev/null
+++ b/db-cleanup.conf
@@ -0,0 +1,21 @@
+#!/hint/bash
+
+# Both PKGPOOLS and SRCPOOLS are relative to `config:FTP_BASE`.
+
+# Directories where packages are shared between repos
+PKGPOOLS=(
+	pool/parabola    # Parabola GNU/Linux-libre
+	pool/alarm       # Arch Linux ARM
+	pool/archlinux32 # Arch Linux 32-bits
+	pool/packages    # Arch Linux (project)
+	pool/community   # Arch Linux (community)
+)
+
+# Directories where sources are stored
+SRCPOOLS=(
+	sources/parabola    # Parabola GNU/Linux-libre
+	sources/alarm       # Arch Linux ARM
+	sources/archlinux32 # Arch Linux 32-bits
+	sources/packages    # Arch Linux (project)
+	sources/community   # Arch Linux (community)
+)
diff --git a/systemd/db-cleanup.service b/systemd/db-cleanup.service
new file mode 100644
index 0000000..669f87e
--- /dev/null
+++ b/systemd/db-cleanup.service
@@ -0,0 +1,7 @@
+[Unit]
+Description=Clean up old files
+
+[Service]
+Type=oneshot
+User=repo
+ExecStart=/opt/dbscripts/cron-jobs/db-cleanup
diff --git a/systemd/db-cleanup.timer b/systemd/db-cleanup.timer
new file mode 100644
index 0000000..8a1c5da
--- /dev/null
+++ b/systemd/db-cleanup.timer
@@ -0,0 +1,9 @@
+[Unit]
+Description=Periodic cleaning of old packages
+
+[Timer]
+OnCalendar=Sun,Tue,Thu,Sat *-*-* 14:13:00
+Persistent=true
+
+[Install]
+WantedBy=timers.target