blancco-init: auto-save reports to the SMB share in process=auto

process=auto erases unattended but Albus forces auto_backup off outside
workflow/manual, so every report needed a manual Save to reach the
share. Debug SSH on a live 7.20 run showed Albus does keep each report
at /reports/<document_id>.xml ("Report saving success", "CRC verified")
and uploads it to MC, with /mnt/share already mounted.

Install blancco-report-sync as a systemd service in the new root: every
10s copy each finished /reports/*.xml (older than 5s, not yet copied)
to /mnt/share as YYYYMMDD_HHMMSS_report.xml (UTC, as a manual Save
names it), via a .part file and rename. Verified live: the auto-run
report landed as 20261005_164520_report.xml.

playbook/blancco-init.sh

@@ -271,6 +271,59 @@ XEOF
 chmod +x /run/newroot/etc/X11/xinit/xinitrc.d/00-no-screen-blank.sh
 
 mkdir -p /run/newroot/proc /run/newroot/sys /run/newroot/dev /run/newroot/run /run/newroot/tmp
+# Auto-save reports to the SMB share. process=auto never writes the share
+# (Albus forces auto_backup off outside workflow/manual) but does keep each
+# report in /reports; a small service copies them across.
+mkdir -p /run/newroot/usr/local/sbin /run/newroot/etc/systemd/system/multi-user.target.wants
+cat > /run/newroot/usr/local/sbin/blancco-report-sync << 'XEOF'
+#!/bin/bash
+# Copy Blancco reports to the SMB report share.
+#
+# In process=auto Albus saves each report to /reports/<document_id>.xml and
+# uploads it to the cloud, but never writes the share: auto_backup is
+# forced off outside workflow/manual mode. Albus has the share mounted at
+# /mnt/share already, so copy each finished report there, named the way a
+# manual Save names it (UTC YYYYMMDD_HHMMSS_report.xml).
+# Usage: blancco-report-sync [--once]
+DONE=/var/lib/blancco-report-sync
+mkdir -p "$DONE"
+while :; do
+    for f in /reports/*.xml; do
+        [ -s "$f" ] || continue
+        id=$(basename "$f" .xml)
+        [ -e "$DONE/$id" ] && continue
+        # Leave files Albus may still be writing.
+        [ $(( $(date +%s) - $(stat -c %Y "$f") )) -ge 5 ] || continue
+        mountpoint -q /mnt/share || continue
+        name="$(date -u -r "$f" +%Y%m%d_%H%M%S)_report.xml"
+        if cp "$f" "/mnt/share/$name.part" && mv -f "/mnt/share/$name.part" "/mnt/share/$name"; then
+            touch "$DONE/$id"
+            echo "copied $id to $name"
+        else
+            rm -f "/mnt/share/$name.part"
+            echo "copy of $id failed, will retry"
+        fi
+    done
+    [ "$1" = "--once" ] && exit 0
+    sleep 10
+done
+XEOF
+chmod 755 /run/newroot/usr/local/sbin/blancco-report-sync
+cat > /run/newroot/etc/systemd/system/blancco-report-sync.service << 'XEOF'
+[Unit]
+Description=Copy Blancco reports to the SMB report share
+
+[Service]
+ExecStart=/usr/local/sbin/blancco-report-sync
+Restart=always
+RestartSec=10
+
+[Install]
+WantedBy=multi-user.target
+XEOF
+ln -sf /etc/systemd/system/blancco-report-sync.service \
+    /run/newroot/etc/systemd/system/multi-user.target.wants/blancco-report-sync.service
+
 # Debug boots only: Blancco's own debug mode (/etc/debug.environment makes
 # isDebugMode true, so 2_sshd starts sshd) plus an SSH key for root.
 # Root is locked and sshd is key-only, so without the key there is no way