3
0
forked from jakub/ansible

Refactor stack health check script to improve readability and structure, ensuring directory existence and enhancing validation for Docker, Nextcloud, database, and Redis connectivity.

This commit is contained in:
fencl
2025-10-03 14:25:29 +02:00
parent 6fbb3fb088
commit bbc27cb2f6

View File

@@ -8,14 +8,23 @@
health_script_path: /data/compose/nextcloud/stack-health.sh
tasks:
- name: Install stack-health.sh (inline content)
- name: Ensure target directory exists
ansible.builtin.file:
path: "{{ health_script_path | dirname }}"
state: directory
mode: '0755'
- name: Install stack-health.sh (inline content, Jinja disabled via raw)
ansible.builtin.copy:
dest: "{{ health_script_path }}"
mode: '0755'
# Comments in English: keep the script inline so we don't depend on repo paths
content: |
# Comments in English: wrap the whole script with {% raw %} ... {% endraw %} so Jinja doesn't eat {{...}} in bash.
content: |-
{% raw %}
#!/usr/bin/env bash
set -euo pipefail
# --- config ---
NC_CONTAINER="${NC_CONTAINER:-nextcloud}"
DB_CONTAINER="${DB_CONTAINER:-nextcloud-db}"
REDIS_CONTAINER="${REDIS_CONTAINER:-redis}"
@@ -23,66 +32,146 @@
NC_HTTP="${NC_HTTP:-http://127.0.0.1:8080/status.php}"
COLLA_DISCOVERY="${COLLA_DISCOVERY:-http://127.0.0.1:9980/hosting/discovery}"
NC_PUBLIC_URL="${NC_PUBLIC_URL:-}"
pass(){ echo -e "✅ $*"; } ; fail(){ echo -e "❌ $*"; EXIT=1; }
EXIT=0; TMP_DIR="$(mktemp -d)"; trap 'rm -rf "$TMP_DIR"' EXIT
pass(){ echo -e "✅ $*"; }
fail(){ echo -e "❌ $*"; EXIT=1; }
EXIT=0
TMP_DIR="$(mktemp -d)"; trap 'rm -rf "$TMP_DIR"' EXIT
# 0) Docker reachable?
docker ps >/dev/null 2>&1 || { fail "Docker not accessible"; echo "----------------------------------------"; echo "ONE OR MORE CHECKS FAILED ❌"; exit "$EXIT"; }
# 1) Containers running?
for c in "$NC_CONTAINER" "$DB_CONTAINER" "$REDIS_CONTAINER" "$COLLA_CONTAINER"; do
if docker ps --format '{{.Names}}' | grep -qx "$c"; then pass "container '$c' is running"; else fail "container '$c' NOT running"; fi
if docker ps --format '{{.Names}}' | grep -qx "$c"; then
pass "container '$c' is running"
else
fail "container '$c' NOT running"
fi
done
# 1b) NC & DB share at least one network?
get_nets(){ docker inspect -f '{{range $k,$v := .NetworkSettings.Networks}}{{printf "%s " $k}}{{end}}' "$1" 2>/dev/null; }
NC_NETS="$(get_nets "$NC_CONTAINER" | xargs || true)"
DB_NETS="$(get_nets "$DB_CONTAINER" | xargs || true)"
COMMON_NET="$(printf "%s\n" $NC_NETS $DB_NETS | tr ' ' '\n' | grep -v '^$' | sort | uniq -d || true)"
[ -n "$COMMON_NET" ] && pass "Nextcloud & DB share network: ${COMMON_NET}" || fail "Nextcloud & DB do NOT share a network (NC: [$NC_NETS], DB: [$DB_NETS])"
if [ -n "$COMMON_NET" ]; then
pass "Nextcloud & DB share network: ${COMMON_NET}"
else
fail "Nextcloud & DB do NOT share a network (NC: [$NC_NETS], DB: [$DB_NETS])"
fi
# 2) OCC status
if docker exec -u www-data "$NC_CONTAINER" php occ status >"$TMP_DIR/occ_status.txt" 2>"$TMP_DIR/occ_status.err"; then
cat "$TMP_DIR/occ_status.txt"
grep -q "installed: true" "$TMP_DIR/occ_status.txt" && pass "Nextcloud installed: true" || fail "Nextcloud installed: false"
grep -q "installed: true" "$TMP_DIR/occ_status.txt" && pass "Nextcloud installed: true" || fail "Nextcloud installed: false"
grep -q "maintenance: false" "$TMP_DIR/occ_status.txt" && pass "Nextcloud maintenance: false" || fail "Nextcloud maintenance is ON"
else fail "occ status failed: $(cat "$TMP_DIR/occ_status.err" || true)"; fi
else
fail "occ status failed: $(cat "$TMP_DIR/occ_status.err" || true)"
fi
# 3) NC -> DB connectivity via PDO from Nextcloud config.php
DB_TEST_PHP="
require \"/var/www/html/lib/base.php\";
\$c=\\OC::\$server->getConfig();
\$h=\$c->getSystemValue(\"dbhost\"); \$d=\$c->getSystemValue(\"dbname\");
\$u=\$c->getSystemValue(\"dbuser\"); \$p=\$c->getSystemValue(\"dbpassword\");
\$h=\$c->getSystemValue(\"dbhost\");
\$d=\$c->getSystemValue(\"dbname\");
\$u=\$c->getSystemValue(\"dbuser\");
\$p=\$c->getSystemValue(\"dbpassword\");
if(!is_string(\$h) || \$h===''){ fwrite(STDERR, \"Invalid dbhost\\n\"); exit(2); }
\$dsn=''; if (\$h[0]==='/' || strpos(\$h,'localhost:/')===0 || strpos(\$h,'127.0.0.1:/')===0) {
\$dsn='';
if (\$h[0]==='/' || strpos(\$h,'localhost:/')===0 || strpos(\$h,'127.0.0.1:/')===0) {
if (strpos(\$h,':/')!==false) { \$parts=explode(':/', \$h, 2); \$sock='/'.\$parts[1]; } else { \$sock=\$h; }
\$dsn='mysql:unix_socket='.\$sock.';dbname='.\$d.';charset=utf8mb4';
} else {
\$host=\$h; \$port=null; if (strpos(\$h,':')!==false) { list(\$host,\$pp)=explode(':',\$h,2); if (ctype_digit(\$pp)) { \$port=(int)\$pp; } }
\$host=\$h; \$port=null;
if (strpos(\$h,':')!==false) { list(\$host,\$pp)=explode(':',\$h,2); if (ctype_digit(\$pp)) { \$port=(int)\$pp; } }
\$dsn='mysql:host='.\$host.';'; if (\$port) { \$dsn.='port='.\$port.';'; }
\$dsn.='dbname='.\$d.';charset=utf8mb4';
}
try { \$pdo=new PDO(\$dsn,\$u,\$p,[PDO::ATTR_TIMEOUT=>2,PDO::ATTR_ERRMODE=>PDO::ERRMODE_EXCEPTION]); \$pdo->query('SELECT 1'); echo \"DB connect OK\\n\"; }
catch (Throwable \$e) { fwrite(STDERR, 'PDO error: '.\$e->getMessage().\"\\n\"); exit(2); }
try {
\$pdo=new PDO(\$dsn,\$u,\$p,[PDO::ATTR_TIMEOUT=>2,PDO::ATTR_ERRMODE=>PDO::ERRMODE_EXCEPTION]);
\$pdo->query('SELECT 1');
echo \"DB connect OK\\n\";
} catch (Throwable \$e) {
fwrite(STDERR, 'PDO error: '.\$e->getMessage().\"\\n\"); exit(2);
}
"
if docker exec -u www-data "$NC_CONTAINER" php -r "$DB_TEST_PHP" >"$TMP_DIR/db_ping.out" 2>"$TMP_DIR/db_ping.err"; then
pass "Nextcloud → DB connection: OK ($(tr -d "\r" <"$TMP_DIR/db_ping.out"))"
else fail "Nextcloud → DB connection FAILED: $(tr -d "\r" <"$TMP_DIR/db_ping.err")"; fi
else
fail "Nextcloud → DB connection FAILED: $(tr -d "\r" <"$TMP_DIR/db_ping.err")"
fi
# 4) DB readiness inside container
if docker exec "$DB_CONTAINER" sh -c 'mariadb-admin ping -h 127.0.0.1 --silent 2>/dev/null || mysqladmin ping -h 127.0.0.1 --silent 2>/dev/null'; then
pass "MariaDB is ready for connections"; else fail "MariaDB readiness check FAILED"; fi
pass "MariaDB is ready for connections"
else
fail "MariaDB readiness check FAILED"
fi
# 5) NC -> Redis
REDIS_TEST_PHP='
require "/var/www/html/lib/base.php";
$c=\OC::$server->getConfig();
$rc=$c->getSystemValue("redis", []);
$host = is_array($rc) && isset($rc["host"]) ? $rc["host"] : (getenv("REDIS_HOST") ?: "redis");
$port = is_array($rc) && isset($rc["port"]) ? (int)$rc["port"] : 6379;
if (class_exists("Redis")) { $r=new Redis(); try { $r->connect($host,$port,1.0); echo $r->ping()."\n"; exit(0); }
catch (Throwable $e) { fwrite(STDERR, "Redis ext error: ".$e->getMessage()."\n"); exit(2); } }
else { $e=$s=0; $fp=@fsockopen($host,$port,$e,$s,1.0);
if (class_exists("Redis")) {
$r=new Redis();
try { $r->connect($host,$port,1.0); echo $r->ping()."\n"; exit(0); }
catch (Throwable $e) { fwrite(STDERR, "Redis ext error: ".$e->getMessage()."\n"); exit(2); }
} else {
$e=$s=0; $fp=@fsockopen($host,$port,$e,$s,1.0);
if($fp){ echo "PONG (tcp ok)\n"; fclose($fp); exit(0); }
fwrite(STDERR,"No php-redis ext and TCP connect failed\n"); exit(2); }'
fwrite(STDERR,"No php-redis ext and TCP connect failed\n"); exit(2);
}
'
if docker exec -u www-data "$NC_CONTAINER" php -r "$REDIS_TEST_PHP" >"$TMP_DIR/redis_ping.out" 2>"$TMP_DIR/redis_ping.err"; then
pass "Nextcloud → Redis: OK ($(tr -d "\r" <"$TMP_DIR/redis_ping.out"))"
else fail "Nextcloud → Redis FAILED: $(tr -d "\r" <"$TMP_DIR/redis_ping.err")"; fi
else
fail "Nextcloud → Redis FAILED: $(tr -d "\r" <"$TMP_DIR/redis_ping.err")"
fi
# 6) Redis local ping
if docker exec "$REDIS_CONTAINER" sh -c 'redis-cli -h 127.0.0.1 ping' >"$TMP_DIR/redis_cli.out" 2>"$TMP_DIR/redis_cli.err"; then
grep -q "PONG" "$TMP_DIR/redis_cli.out" && pass "Redis container responds to PING" || fail "Redis cli ping did not return PONG"
else fail "redis-cli ping failed: $(cat "$TMP_DIR/redis_cli.err" || true)"; fi
if curl -fsS "$NC_HTTP" -o "$TMP_DIR/nc_http.json"; then pass "HTTP $NC_HTTP reachable"; head -c 300 "$TMP_DIR/nc_http.json"; echo; else fail "HTTP $NC_HTTP unreachable"; fi
if [ -n "$NC_PUBLIC_URL" ]; then curl -fsSIL "$NC_PUBLIC_URL" >/dev/null && pass "Public URL reachable: $NC_PUBLIC_URL" || fail "Public URL unreachable: $NC_PUBLIC_URL"; fi
if curl -fsS "$COLLA_DISCOVERY" -o "$TMP_DIR/colla_disc.xml"; then pass "Collabora discovery reachable ($COLLA_DISCOVERY)"; grep -m1 -o "<wopi-discovery[^>]*>" "$TMP_DIR/colla_disc.xml" || true; else fail "Collabora discovery unreachable: $COLLA_DISCOVERY"; fi
if [ "$EXIT" -eq 0 ]; then echo "----------------------------------------"; echo "ALL CHECKS PASSED ✅"; else echo "----------------------------------------"; echo "ONE OR MORE CHECKS FAILED ❌ (see above)"; fi
if grep -q "PONG" "$TMP_DIR/redis_cli.out"; then pass "Redis container responds to PING"; else fail "Redis cli ping did not return PONG"; fi
else
fail "redis-cli ping failed: $(cat "$TMP_DIR/redis_cli.err" || true)"
fi
# 7) HTTP status.php
if curl -fsS "$NC_HTTP" -o "$TMP_DIR/nc_http.json"; then
pass "HTTP $NC_HTTP reachable"
head -c 300 "$TMP_DIR/nc_http.json"; echo
else
fail "HTTP $NC_HTTP unreachable"
fi
# 8) Optional public URL
if [ -n "$NC_PUBLIC_URL" ]; then
if curl -fsSIL "$NC_PUBLIC_URL" >/dev/null; then pass "Public URL reachable: $NC_PUBLIC_URL"; else fail "Public URL unreachable: $NC_PUBLIC_URL"; fi
fi
# 9) Collabora discovery
if curl -fsS "$COLLA_DISCOVERY" -o "$TMP_DIR/colla_disc.xml"; then
pass "Collabora discovery reachable ($COLLA_DISCOVERY)"
grep -m1 -o "<wopi-discovery[^>]*>" "$TMP_DIR/colla_disc.xml" || true
else
fail "Collabora discovery unreachable: $COLLA_DISCOVERY"
fi
# summary
if [ "$EXIT" -eq 0 ]; then
echo "----------------------------------------"
echo "ALL CHECKS PASSED ✅"
else
echo "----------------------------------------"
echo "ONE OR MORE CHECKS FAILED ❌ (see above)"
fi
exit "$EXIT"
{% endraw %}
- name: Run stack-health.sh
ansible.builtin.shell: "{{ health_script_path }}"
@@ -94,7 +183,7 @@
ansible.builtin.debug:
msg: "{{ health.stdout | default('no stdout') }}"
- name: Fail if checks failed
- name: Fail if checks failed (rc != 0)
ansible.builtin.fail:
msg: "Health checks failed"
when: health.rc != 0