GeezerVison

This repo provides simple sync job control.

Default database name: mirrors

Columns:

• index: index, primary key
• id: name of the mirror, also unique indentifier
• state: state of the mirror, in {'error','sync','done'}
• lastSyncTime: timestamp (in milliseconds) of last sync
• nextSyncTime: timestamp (in milliseconds) of next sync
• diskUsage: disk usage in string form
• logPath: where the lastest logfile is

to create table:

CREATE TABLE `status`(
    `index` INT UNSIGNED AUTO_INCREMENT ,
    `id` VARCHAR(50) NOT NULL,
    `state` VARCHAR(20) NOT NULL,
    `lastSyncTime` BIGINT UNSIGNED DEFAULT 0,
    `nextSyncTime` BIGINT UNSIGNED DEFAULT 0,
    `diskUsage` VARCHAR(30) DEFAULT '',
    `logPath` VARCHAR(100) DEFAULT '',
    `laststate` VARCHAR(20) NOT NULL,
    PRIMARY KEY (`index`) ) ENGINE=InnoDB DEFAULT CHARSET=utf8;

• GET /status

  returns current status in JSON

  an example of response:

  [
      {
          "id": "manjaro",
          "state": "done",
          "lastSyncTime":1615911140966,
          "nextSyncTime":1615997400000,
          "diskUsage":"100.43G"
      }
  ]
  

CAUTION: Following APIs should be accessed by admin only.

• GET /stop/$id

  stop the job with id=$id

  returns a JSON

  // success
  {"status":"ok"}
  
  // failure
  {"status":"failed","reason":"?"}
  

• GET /start/$id

  start the job with id=$id

  returns a JSON

  // success
  {"status":"ok"}
  
  // failure
  {"status":"failed","reason":"?"}
  

• GET /log/$id

  return the job's latest log

Configuration file is ./config.json by default. You can assign custom configuration path by adding -c <path> in command-line.

Here is an example of full configuration

{
    "mirrors": [
        {  
            "id": "manjaro",
            "schedule": "10 0 * * *",
            "provider": "rsync",
            "params": {
                "url": "rsync://mirror.yandex.ru/mirrors/manjaro/"
            },
            "metadata": {
                "name": "Manjaro",
                "url": "/manjaro/"
            },
            "file-custom": [
                "alias /abc/def/;"
            ]
        },
        {  
            "id": "opensuse",
            "schedule": "0 */8 * * *",
            "provider": "rsync",
            "params": {
                "url": "rsync://fr2.rpmfind.net/linux/opensuse/",
                "exclude": "{history,ports,source}",
                "dest": "/data-tmp/opensuse"
            }
        },
        {
            "id": "custom",
            "schedule": "10 0 * * *",
            "provider": "shell",
            "params": {
                "cmd": "node ./update_custom.js"
            }
        }
    ],
    "concurrency": 2,
    "log_dir": "/data/log/sync/",
    "repo_dir": "/data/repos/",
    "rsync": {
        "command": "rsync -avHh --delete --delete-after --delay-updates --safe-links --stats --no-o --no-g"
    },
    "mysql": {
        "host": "127.0.0.1",
        "user": "sync",
        "password": "123456",
        "database": "sync"
    },
    "port":3000
}

Attention: metadata in a mirror is used to render mirror list.

here is an example:

"metadata": {
    "url": "/ubuntu/",
    "name": "Ubuntu",
    "help": "Ubuntu",
    "describe": "A Ubuntu mirror"
}

The describe will show when your mouse hover on the link.

The help should satisfy the requirement in website/src/components/help/Ubuntu.vue currently

Be aware:

• mirrors stores the list of sync jobs, in which schedule is a cron-style string indicating frequency (more info). We implemented two providers, rsync and shell, you can use them as the way shown above.
• mysql stores the database connection params. See mysql for more info.
• concurrency limits the max concurrency of run_queue.

Typical way:

1. clone the repo
2. modify the configuration
3. node index.js

You can assign configuration file by node index.js -c <path> or node index.js --conf <path>.

We also provides systemd service.

Remember to modify WorkingDirectory & ExecStart in backend.service to correct path before use

# cp backend.service /etc/systemd/system
# systemctl start backend.service

• manual kill
• manual start sync job
• status + disk usage (of *** job)
• hot reload config (get/update)
• restart(kill all running job & restart)
• rsync provider
• add dynamic proxy function => tip: use nginx_http_dyups_module in nginx and dynamicly change upstream
• may need a slave sync controller run on the second machine to avoid sync the same mirror at a same time