feat: add cleanup to autopromoter (#393)

- ensure the autopromoter removes hardlinks/replicas for repos older
  than the current promoted monthly
- this is to reduce MDS load for ceph, as hardlinks require memory

Reviewed-on: #393

site/profiles/templates/reposync/autopromoter.erb

@@ -1,33 +1,72 @@
 #!/usr/bin/env bash
+set -euo pipefail
+
+: "${DRY_RUN:=}"   # set DRY_RUN=1 to preview deletions
+
+log() { printf '[%(%F %T)T] %s\n' -1 "$*" >&2; }
 
 # Function to create symlink for snapshots
 create_symlink() {
-  local osname="$1"
-  local release="$2"
-  local repository="$3"
-  local basepath="$4"
-  local label="$5" # 'monthly', 'weekly', or 'daily'
-  local date_format="$6" # Date format for finding the snapshot
+  local osname="$1"     # e.g. almalinux
+  local release="$2"    # e.g. 9.6
+  local repository="$3" # e.g. appstream
+  local basepath="$4"   # e.g. /shared/apps/packagerepo
+  local label="$5"      # monthly|weekly|daily
+  local date_value="$6" # YYYYMMDD for the snapshot to promote
 
-  # The path where snapshots are stored
-  local snap_path="${basepath}/snap/${osname}/${release}/${repository}-${date_format}"
-
-  # The target path for the symlink
+  local snap_path="${basepath}/snap/${osname}/${release}/${repository}-${date_value}"
   local symlink_target="${basepath}/snap/${osname}/${release}/${repository}-${label}"
 
-  # Check if the source directory exists
   if [[ -d "$snap_path" ]]; then
-    # Create the symlink, overwrite if it already exists
     ln -sfn "$snap_path" "$symlink_target"
-    echo "Symlink created for $snap_path -> $symlink_target"
+    log "Symlink set: ${symlink_target} -> ${snap_path}"
+    return 0
   else
-    echo "Snapshot path does not exist: $snap_path"
+    log "Snapshot path does not exist: $snap_path"
     return 1
   fi
 }
 
+# Cleanup snapshot directories older than a cutoff date (YYYYMMDD)
+cleanup_older_than_monthly() {
+  local osname="$1"
+  local release="$2"
+  local repository="$3"
+  local basepath="$4"
+  local cutoff_date="$5"  # newly promoted monthly date, YYYYMMDD
+
+  local root="${basepath}/snap/${osname}/${release}"
+  local pattern="${repository}-"[0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]
+
+  shopt -s nullglob
+  for dir in "${root}"/${pattern}; do
+    # Skip symlinks and non-directories
+    [[ -L "$dir" ]] && continue
+    [[ -d "$dir" ]] || continue
+
+    # Extract the date suffix
+    local name base datepart
+    name="$(basename -- "$dir")"
+    base="${name%-????????}"     # remove last 8 chars
+    datepart="${name#${base}-}"  # after "<repo>-"
+
+    # Only act on valid YYYYMMDD dates
+    if [[ "$datepart" =~ ^[0-9]{8}$ ]]; then
+      if [[ "$datepart" -lt "$cutoff_date" ]]; then
+        if [[ -n "${DRY_RUN}" ]]; then
+          log "[DRY_RUN] Would remove: $dir"
+        else
+          log "Removing: $dir"
+          rm -rf --one-file-system -- "$dir"
+        fi
+      fi
+    fi
+  done
+  shopt -u nullglob
+}
+
 # Determine which snapshot to promote based on the passed argument
-case "$1" in
+case "${1:-}" in
   monthly)
     promote_date=$(date --date="$(date +%Y%m01) -1 month" +%Y%m%d)
     ;;
@@ -43,11 +82,26 @@ case "$1" in
     ;;
 esac
 
-# Call the function with appropriate arguments
-# Iterate over the repositories to create symlinks for each
+label="$1"
+
+# Process each repo conf
 for conf in /etc/reposync/conf.d/*.conf; do
+  # shellcheck disable=SC1090
   source "$conf"
 
-  # Create symlink based on the provided argument
-  create_symlink "$OSNAME" "$RELEASE" "$REPOSITORY" "$BASEPATH" "$1" "$promote_date"
+  # Expecting these vars in the conf: OSNAME, RELEASE, REPOSITORY, BASEPATH
+  : "${OSNAME:?missing OSNAME in $conf}"
+  : "${RELEASE:?missing RELEASE in $conf}"
+  : "${REPOSITORY:?missing REPOSITORY in $conf}"
+  : "${BASEPATH:?missing BASEPATH in $conf}"
+
+  # 1) Promote the symlink for this label
+  if create_symlink "$OSNAME" "$RELEASE" "$REPOSITORY" "$BASEPATH" "$label" "$promote_date"; then
+    # 2) If monthly, clean up older snapshot dirs for this repo
+    if [[ "$label" == "monthly" ]]; then
+      cleanup_older_than_monthly "$OSNAME" "$RELEASE" "$REPOSITORY" "$BASEPATH" "$promote_date"
+    fi
+  else
+    log "Skipping cleanup for ${OSNAME}/${RELEASE}/${REPOSITORY} due to missing ${REPOSITORY}-${promote_date}"
+  fi
 done