新增备份监控与定时配置

This commit is contained in:
yml2213
2026-08-16 23:53:50 +08:00
parent 66cb03ebce
commit f67ee12d72
19 changed files with 612 additions and 3 deletions
+83
View File
@@ -0,0 +1,83 @@
#!/usr/bin/env bash
set -euo pipefail
umask 077
# 由宿主机 cron 每分钟调用。调度配置来自后台 system_configs,实际执行仍在宿主机,
# Web 后端不接触 Docker socket、备份载荷或客户端加密密码。
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
# shellcheck source=backup-common.sh
source "${SCRIPT_DIR}/backup-common.sh"
config_value() {
local key="$1" fallback="$2" value
value="$(backup_mysql_root_query "SELECT value FROM system_configs WHERE \`key\` = '${key}' LIMIT 1;" 2>/dev/null || true)"
value="${value//$'\n'/}"
printf '%s' "${value:-${fallback}}"
}
write_job_state() {
local status_dir="$1" type="$2" status="$3" message="$4"
mkdir -p "${status_dir}"
{
printf 'job_type=%q\n' "${type}"
printf 'job_status=%q\n' "${status}"
printf 'job_at=%q\n' "$(date -Iseconds)"
printf 'job_message=%q\n' "${message}"
} > "${status_dir}/last-job.env"
chmod 600 "${status_dir}/last-job.env"
}
run_job() {
local status_dir="$1" type="$2" script="$3"
shift 3
if "${script}" "$@"; then
write_job_state "${status_dir}" "${type}" success '执行成功'
"${SCRIPT_DIR}/backup-status.sh" write || true
return
fi
local code=$?
write_job_state "${status_dir}" "${type}" failed "执行失败(退出码 ${code}"
"${SCRIPT_DIR}/backup-status.sh" write || true
return "${code}"
}
main() {
backup_require_cmd docker
backup_require_cmd flock
local backup_dir status_dir enabled hour minute interval now_hour now_minute
backup_dir="$(backup_require_env BACKUP_DIR)"
status_dir="${BACKUP_STATUS_DIR:-${backup_dir}/status}"
backup_validate_dir "${status_dir}"
backup_lock_or_exit "${backup_dir}/.scheduler.lock"
enabled="$(config_value 'backup.schedule_enabled' 'true')"
hour="$(config_value 'backup.full_hour' '4')"
minute="$(config_value 'backup.full_minute' '30')"
interval="$(config_value 'backup.binlog_interval_minutes' '1')"
[[ "${enabled}" == true || "${enabled}" == false ]] || { backup_error 'backup.schedule_enabled 必须为 true 或 false'; exit 1; }
[[ "${hour}" =~ ^[0-9]+$ && "${hour}" -le 23 && "${minute}" =~ ^[0-9]+$ && "${minute}" -le 59 ]] || {
backup_error '备份全量时间配置无效'
exit 1
}
[[ "${interval}" =~ ^[0-9]+$ && "${interval}" -ge 1 && "${interval}" -le 5 ]] || {
backup_error 'binlog 归档间隔必须为 1-5 分钟'
exit 1
}
if [[ "${enabled}" == false ]]; then
"${SCRIPT_DIR}/backup-status.sh" write || true
exit 0
fi
now_hour="$((10#$(date +%H)))"
now_minute="$((10#$(date +%M)))"
if ((now_minute % interval == 0)); then
run_job "${status_dir}" binlog "${SCRIPT_DIR}/archive-binlog.sh" archive
fi
if ((now_hour == hour && now_minute == minute)); then
run_job "${status_dir}" full "${SCRIPT_DIR}/backup-online.sh" full
fi
"${SCRIPT_DIR}/backup-status.sh" write || true
}
main "$@"
+106
View File
@@ -0,0 +1,106 @@
#!/usr/bin/env bash
set -euo pipefail
umask 077
# 生成供后台只读展示的备份状态摘要。内容不包含密钥、备份载荷或下载地址。
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
# shellcheck source=backup-common.sh
source "${SCRIPT_DIR}/backup-common.sh"
json_escape() {
local value="$1"
value="${value//\\/\\\\}"
value="${value//\"/\\\"}"
value="${value//$'\n'/ }"
value="${value//$'\r'/ }"
printf '%s' "${value}"
}
config_value() {
local key="$1" fallback="$2" value
value="$(backup_mysql_root_query "SELECT value FROM system_configs WHERE \`key\` = '${key}' LIMIT 1;" 2>/dev/null || true)"
value="${value//$'\n'/}"
printf '%s' "${value:-${fallback}}"
}
write_status() {
backup_require_cmd docker
backup_require_cmd ossutil
local backup_dir oss_uri status_dir status_file temp_file enabled hour minute interval
local local_complete remote_complete binlog_state binlog_file binlog_remote job_type job_status job_at job_message health
backup_dir="$(backup_require_env BACKUP_DIR)"
oss_uri="$(backup_require_env BACKUP_OSS_URI)"
status_dir="${BACKUP_STATUS_DIR:-${backup_dir}/status}"
backup_validate_dir "${status_dir}"
mkdir -p "${status_dir}"
chmod 700 "${status_dir}"
status_file="${status_dir}/status.json"
temp_file="${status_dir}/.status-$$.json"
enabled="$(config_value 'backup.schedule_enabled' 'true')"
hour="$(config_value 'backup.full_hour' '4')"
minute="$(config_value 'backup.full_minute' '30')"
interval="$(config_value 'backup.binlog_interval_minutes' '1')"
local_complete="$(find "${backup_dir}/full/daily" -name complete.env -type f -print 2>/dev/null | sort | tail -1 || true)"
remote_complete="$(ossutil ls "${oss_uri}/mysql/full/daily/" -r 2>/dev/null | awk '$NF ~ /\/complete\.env$/ { latest = $NF } END { print latest }')"
binlog_state="$(find "${backup_dir}/binlog-state" -name last-archived.env -type f -print 2>/dev/null | sort | tail -1 || true)"
binlog_file=''
binlog_remote=''
if [[ -n "${binlog_state}" ]]; then
binlog_file="$(awk -F= '$1 == "filename" { print substr($0, index($0, "=") + 1); exit }' "${binlog_state}")"
binlog_remote="$(awk -F= '$1 == "remote_dir" { print substr($0, index($0, "=") + 1); exit }' "${binlog_state}")"
fi
job_type=''
job_status=''
job_at=''
job_message=''
if [[ -f "${status_dir}/last-job.env" ]]; then
# shellcheck disable=SC1090
source "${status_dir}/last-job.env"
fi
if [[ -n "${local_complete}" && -n "${remote_complete}" && -n "${binlog_file}" ]]; then
health='healthy'
else
health='warning'
fi
cat > "${temp_file}" <<EOF
{
"generated_at": "$(date -Iseconds)",
"health": "$(json_escape "${health}")",
"schedule": {
"enabled": "$(json_escape "${enabled}")",
"full_hour": "$(json_escape "${hour}")",
"full_minute": "$(json_escape "${minute}")",
"binlog_interval_minutes": "$(json_escape "${interval}")"
},
"last_full": {
"local_complete": "$(json_escape "${local_complete}")",
"remote_complete": "$(json_escape "${remote_complete}")"
},
"last_binlog": {
"file": "$(json_escape "${binlog_file}")",
"remote_dir": "$(json_escape "${binlog_remote}")"
},
"last_job": {
"type": "$(json_escape "${job_type}")",
"status": "$(json_escape "${job_status}")",
"at": "$(json_escape "${job_at}")",
"message": "$(json_escape "${job_message}")"
}
}
EOF
chmod 640 "${temp_file}"
mv "${temp_file}" "${status_file}"
}
case "${1:-write}" in
write) write_status ;;
*)
printf '用法: %s {write}\n' "$0" >&2
exit 1
;;
esac
+25 -1
View File
@@ -146,7 +146,7 @@ validate_env() {
exit 1
fi
local app_env caddy_domain caddy_email mysql_dsn redis_addr storage_endpoint backend_uid backend_gid backup_dir backup_restore_dir backup_oss_uri backup_user
local app_env caddy_domain caddy_email mysql_dsn redis_addr storage_endpoint backend_uid backend_gid backup_dir backup_status_dir backup_restore_dir backup_oss_uri backup_user
app_env="$(require_env APP_ENV)"
caddy_domain="$(require_env CADDY_DOMAIN)"
caddy_email="$(require_env CADDY_EMAIL)"
@@ -162,6 +162,7 @@ validate_env() {
require_env STORAGE_ACCESS_KEY_ID >/dev/null
require_env STORAGE_SECRET_ACCESS_KEY >/dev/null
backup_dir="$(require_env BACKUP_DIR)"
backup_status_dir="$(require_env BACKUP_STATUS_DIR)"
backup_restore_dir="$(require_env BACKUP_RESTORE_DIR)"
backup_oss_uri="$(require_env BACKUP_OSS_URI)"
require_env BACKUP_PASSPHRASE >/dev/null
@@ -216,6 +217,10 @@ validate_env() {
log_error "BACKUP_DIR 必须是 /data 之下的具体绝对路径,当前值:${backup_dir}"
exit 1
fi
if [[ "${backup_status_dir}" != "${backup_dir}/"* ]]; then
log_error "BACKUP_STATUS_DIR 必须位于 BACKUP_DIR 之下,当前值:${backup_status_dir}"
exit 1
fi
if [[ "${backup_restore_dir}" != /data/* || "${backup_restore_dir}" == "/data" || "${backup_restore_dir}" == "${backup_dir}" ]]; then
log_error "BACKUP_RESTORE_DIR 必须是与 BACKUP_DIR 不同的 /data 子目录,当前值:${backup_restore_dir}"
exit 1
@@ -403,6 +408,24 @@ prepare_log_dir() {
chmod 700 "${BACKEND_LOG_DIR}"
}
prepare_backup_status_dir() {
local backend_uid backend_gid status_dir owner_mismatch
backend_uid="$(require_env BACKEND_UID)"
backend_gid="$(require_env BACKEND_GID)"
status_dir="$(require_env BACKUP_STATUS_DIR)"
mkdir -p "${status_dir}"
owner_mismatch="$(find "${status_dir}" \( ! -uid "${backend_uid}" -o ! -gid "${backend_gid}" \) -print -quit)"
if [[ -n "${owner_mismatch}" ]]; then
if [[ "${EUID}" -ne 0 ]]; then
log_error "备份状态目录存在属主不匹配的文件:${owner_mismatch}"
log_error "请先执行:sudo chown -R ${backend_uid}:${backend_gid} ${status_dir}"
exit 1
fi
chown -R "${backend_uid}:${backend_gid}" "${status_dir}"
fi
chmod 700 "${status_dir}"
}
wait_service_healthy() {
local service="$1"
local label="$2"
@@ -557,6 +580,7 @@ main() {
validate_env
export_caddy_env
prepare_log_dir
prepare_backup_status_dir
# 选号网需在 build caddy 之前准备 show-dist 与 conf.d
prepare_show_site