#!/usr/bin/env bash
# felhom-tenantsync v1.0.0 — the offsite endpoint's per-customer PBS tenancy surface (PBS DR tier
# SLICE 1; spike SPIKE-pbs-tier-provisioning-2026-07-10 §3).
#
# Runs as the SSH forced command for the hub's SECOND `felhom-peersync` key (via sudo — its own
# single sudoers line; the peersync script/key are untouched: one script, one job). JSON on stdin,
# JSON on stdout. Ops:
#
#   {"op":"provision","customer_id":"<id>"}  → ensure namespace <id> (idempotent) → CREATE token
#       felhom@pbs!<id> (an EXISTING token is a hard error, code "token_exists" — re-issue is the
#       explicit path) → dual-grant DatastoreBackup on /datastore/felhom-offsite/<id> to BOTH the
#       user and the token (PBS privsep = intersection) → self-check: list the namespace AS the
#       new token (one regen retry per the spike's transient-403 note; still failing → rollback)
#       → {"status":"ok","token_id","token_secret","fingerprint","datastore","namespace"}
#   {"op":"reissue","customer_id":"<id>"}    → delete-token (its ACLs purge with it — spike) →
#       recreate → re-grant BOTH → self-check → same ok-shape with the FRESH secret.
#   {"op":"fingerprint"}                     → {"status":"ok","fingerprint":"<PBS cert sha256>"}
#
# NO deprovision op in slice 1 — namespace/data deletion is a deliberate, separate decision (the
# offsite-disable precedent: disable never destroys data).
#
# Secret hygiene (load-bearing):
#   - The token secret exists ONLY in memory and in the final stdout JSON — never a file, never
#     stderr (the hub embeds remote stderr in error logs), never an argument.
#   - The transient root@pam admin token (namespace ops are client-side) is held in memory and
#     deleted on EVERY exit path (trap). A leftover from a crashed run is deleted at entry.
#   - All tool stdout is redirected to stderr — the response JSON is the ONLY stdout bytes.
#
# Ordering facts this script encodes (all live-proven in the spike — do not "simplify"):
#   - the token must exist BEFORE its ACL grant (PBS validates the auth-id);
#   - user delete-token PURGES the token's ACLs → re-issue must re-grant;
#   - `user generate-token` has no --output-format → the "value" line is sed-parsed;
#   - proxmox-backup-client needs PBS_FINGERPRINT even against localhost.
set -euo pipefail

DS=felhom-offsite
PBS_USER=felhom@pbs
ADMIN_TOKEN_NAME=tenantsync-admin
REPO_HOST=localhost

err_json() { # code, message → error JSON on stdout, exit 1
  printf '{"status":"error","code":"%s","error":"%s"}\n' "$1" "$2"
  exit 1
}

command -v jq >/dev/null || err_json internal "jq is required"
command -v proxmox-backup-manager >/dev/null || err_json internal "proxmox-backup-manager is required"
command -v proxmox-backup-client >/dev/null || err_json internal "proxmox-backup-client is required"

# 1. Read stdin capped at 64 KiB; validate the envelope BEFORE touching anything.
payload=$(head -c 65536)
[ -n "$payload" ] || err_json bad_request "empty payload"
jq -e 'type == "object" and (.op | type) == "string"' >/dev/null 2>&1 <<<"$payload" \
  || err_json bad_request "payload must be a JSON object with a string op"
OP=$(jq -r '.op' <<<"$payload")

export PBS_FINGERPRINT
PBS_FINGERPRINT=$(proxmox-backup-manager cert info | awk '/Fingerprint/{print $3}' | head -1)
[ -n "$PBS_FINGERPRINT" ] || err_json internal "could not read the PBS cert fingerprint"

if [ "$OP" = "fingerprint" ]; then
  printf '{"status":"ok","fingerprint":"%s"}\n' "$PBS_FINGERPRINT"
  exit 0
fi

case "$OP" in provision|reissue) ;; *) err_json bad_request "unknown op" ;; esac

# customer_id → the namespace AND the token name. Conservative charset (PBS ns + token grammar,
# no leading dash/dot so it can never parse as an option).
CID=$(jq -r '.customer_id // ""' <<<"$payload")
[[ "$CID" =~ ^[A-Za-z0-9][A-Za-z0-9_.-]{0,30}$ ]] \
  || err_json bad_request "customer_id must match ^[A-Za-z0-9][A-Za-z0-9_.-]{0,30}\$"
TOKEN_ID="$PBS_USER!$CID"

log() { echo "felhom-tenantsync: $*" >&2; }

# 2. Transient admin token for the client-side namespace ops. Deleted on every exit; a leftover
#    from a crashed run is cleared first (generate-token fails on an existing name).
proxmox-backup-manager user delete-token root@pam "$ADMIN_TOKEN_NAME" >&2 2>/dev/null || true
ADM=$(proxmox-backup-manager user generate-token root@pam "$ADMIN_TOKEN_NAME" \
  | sed -n 's/.*"value": "\([^"]*\)".*/\1/p')
[ -n "$ADM" ] || err_json internal "admin token generation failed"
cleanup_admin() {
  proxmox-backup-manager user delete-token root@pam "$ADMIN_TOKEN_NAME" >&2 2>/dev/null || true
  proxmox-backup-manager acl update "/datastore/$DS" DatastoreAdmin \
    --auth-id "root@pam!$ADMIN_TOKEN_NAME" --delete >&2 2>/dev/null || true
}
trap cleanup_admin EXIT
proxmox-backup-manager acl update "/datastore/$DS" DatastoreAdmin \
  --auth-id "root@pam!$ADMIN_TOKEN_NAME" >&2

ADMIN_REPO="root@pam!$ADMIN_TOKEN_NAME@$REPO_HOST:$DS"

# 3. Ensure the shared user + the namespace (both idempotent).
if ! proxmox-backup-manager user list --output-format json | jq -e --arg u "$PBS_USER" \
    'any(.[]; .userid == $u)' >/dev/null; then
  proxmox-backup-manager user create "$PBS_USER" --comment 'offsite tenancy' >&2
  log "created user $PBS_USER"
fi
# NOTE: the CLIENT wraps json output as {"data":[...]} (live-proven on ep0); the MANAGER commands
# return bare arrays. `(.data // .)` handles both.
if ! PBS_PASSWORD="$ADM" proxmox-backup-client namespace list --repository "$ADMIN_REPO" \
    --output-format json | jq -e --arg ns "$CID" '(.data // .) | any(.[]; .ns == $ns)' >/dev/null; then
  PBS_PASSWORD="$ADM" proxmox-backup-client namespace create "$CID" --repository "$ADMIN_REPO" >&2
  log "created namespace $CID"
fi

token_exists() {
  proxmox-backup-manager user list-tokens "$PBS_USER" --output-format json \
    | jq -e --arg t "$TOKEN_ID" 'any(.[]; .tokenid == $t)' >/dev/null
}

# gen_token → SECRET on stdout of this function only (command substitution), nothing persisted.
gen_token() {
  proxmox-backup-manager user generate-token "$PBS_USER" "$CID" \
    | sed -n 's/.*"value": "\([^"]*\)".*/\1/p'
}

grant_both() { # dual-grant on the NAMESPACE ACL path (never /ns/<ns> — §4a gotcha)
  proxmox-backup-manager acl update "/datastore/$DS/$CID" DatastoreBackup --auth-id "$PBS_USER" >&2
  proxmox-backup-manager acl update "/datastore/$DS/$CID" DatastoreBackup --auth-id "$TOKEN_ID" >&2
}

self_check() { # own-namespace list AS the new token; secret via env, never argv/stderr
  PBS_PASSWORD="$1" proxmox-backup-client snapshot list --ns "$CID" \
    --repository "$TOKEN_ID@$REPO_HOST:$DS" >/dev/null
}

rollback_token() {
  proxmox-backup-manager user delete-token "$PBS_USER" "$CID" >&2 2>/dev/null || true
  proxmox-backup-manager acl update "/datastore/$DS/$CID" DatastoreBackup \
    --auth-id "$PBS_USER" --delete >&2 2>/dev/null || true
}

if [ "$OP" = "provision" ]; then
  if token_exists; then
    err_json token_exists "token $TOKEN_ID already exists — use the reissue op (explicit re-key)"
  fi
else # reissue: delete-token purges its ACLs; recreate + re-grant below
  proxmox-backup-manager user delete-token "$PBS_USER" "$CID" >&2 2>/dev/null || true
  log "reissue: old token deleted (ACLs purged with it)"
fi

SECRET=$(gen_token)
[ -n "$SECRET" ] || err_json internal "token generation returned no value"
grant_both

if ! self_check "$SECRET"; then
  # The spike's transient-403 note: one delete+regen+re-grant retry, then rollback + fail.
  log "self-check failed — regenerating once (spike transient-403 note)"
  proxmox-backup-manager user delete-token "$PBS_USER" "$CID" >&2 2>/dev/null || true
  SECRET=$(gen_token)
  [ -n "$SECRET" ] || { rollback_token; err_json internal "token regeneration returned no value"; }
  grant_both
  if ! self_check "$SECRET"; then
    rollback_token
    err_json self_check_failed "own-namespace list as $TOKEN_ID failed twice — rolled back"
  fi
fi

log "$OP ok: ns=$CID token=$TOKEN_ID (secret rides stdout only)"
jq -cn --arg tid "$TOKEN_ID" --arg sec "$SECRET" --arg fp "$PBS_FINGERPRINT" \
  --arg ds "$DS" --arg ns "$CID" \
  '{"status":"ok","token_id":$tid,"token_secret":$sec,"fingerprint":$fp,"datastore":$ds,"namespace":$ns}'