Trouble setting up incremental backup in MongoDB Cluster - Docker

MongoDB Percona Backup for MongoDB
1

Hello everyone,

I’m currently working on setting up a new MongoDB cluster with 3 nodes for development and testing purposes. All three nodes are hosted on the same machine. I’m using Docker to run this MongoDB cluster along with containers for percona-backup-mongodb. Below is my docker-compose.yml file:

version: "3"
services:
  rs101:
    image: percona/percona-server-mongodb:6.0
    container_name: rs101
    hostname: rs101
    ports:
      - "27017:27017"
    networks:
      - db-network
    command: "--port=27017 --replSet rs"

  rs102:
    image: percona/percona-server-mongodb:6.0
    container_name: rs102
    hostname: rs102
    ports:
      - "28017:28017"
    networks:
      - db-network
    command: "--port=28017 --replSet rs"

  rs103:
    image: percona/percona-server-mongodb:6.0
    container_name: rs103
    hostname: rs103
    ports:
      - "29017:29017"
    networks:
      - db-network
    command: "--port=29017 --replSet rs"

  percona-backup-mongod-1:
    container_name: pbm1
    networks:
      - db-network
    environment:
      - PBM_MONGODB_URI=mongodb://pbmuser:secretpwd@rs101:27017
    image: percona/percona-backup-mongodb:latest

  percona-backup-mongodb-2:
    container_name: pbm2
    networks:
      - db-network
    environment:
      - PBM_MONGODB_URI=mongodb://pbmuser:secretpwd@rs102:28017
    image: percona/percona-backup-mongodb:latest

  percona-backup-mongodb-3:
    container_name: pbm3
    networks:
      - db-network
    environment:
      - PBM_MONGODB_URI=mongodb://pbmuser:secretpwd@rs103:29017
    image: percona/percona-backup-mongodb:latest

  rs-init:
    image: percona/percona-server-mongodb:6.0
    container_name: rs-init
    restart: "no"
    networks:
      - db-network
    depends_on:
      - rs101
      - rs102
      - rs103
    command: >
      mongosh --host rs101:27017 --eval
      '
      config = {
      "_id" : "rs",
      "members" : [
        {
          "_id" : 0,
          "host" : "rs101:27017"
        },
        {
          "_id" : 1,
          "host" : "rs102:28017"
        },
        {
          "_id" : 2,
          "host" : "rs103:29017"
        }
      ]
      };
      rs.initiate(config);
      '
networks:
  db-network:
    driver: bridge

At this stage, I’ve been successful in creating logical snapshot backups, which I store in a MinIO bucket:

$ pbm list
Backup snapshots:
  2024-04-03T10:10:53Z <logical> [restore_to_time: 2024-04-03T10:11:04Z]
  2024-04-03T11:49:10Z <logical> [restore_to_time: 2024-04-03T11:49:23Z]
  2024-04-03T11:54:49Z <logical> [restore_to_time: 2024-04-03T11:55:00Z]
  2024-04-03T11:56:48Z <logical> [restore_to_time: 2024-04-03T11:57:00Z]
  2024-04-03T12:31:19Z <logical> [restore_to_time: 2024-04-03T12:31:25Z]

PITR <on>:

However, I’m encountering difficulties in implementing incremental backups:

$ pbm backup --type incremental --base
Starting backup '2024-04-03T13:00:59Z'....Backup '2024-04-03T13:00:59Z' to remote store 's3://xxxxxxxxxxxx/mongo-backup' has started

$ pbm logs -s D -t 1000
2024-04-03T13:01:00Z I [rs/rs101:27017] got command backup [name: 2024-04-03T13:00:59Z, compression: s2 (level: default)] <ts: 1712149259>
2024-04-03T13:01:00Z I [rs/rs102:28017] got command backup [name: 2024-04-03T13:00:59Z, compression: s2 (level: default)] <ts: 1712149259>
2024-04-03T13:01:00Z I [rs/rs101:27017] got epoch {1712148849 3}
2024-04-03T13:01:00Z I [rs/rs102:28017] got epoch {1712148849 3}
2024-04-03T13:01:00Z I [rs/rs103:29017] got command backup [name: 2024-04-03T13:00:59Z, compression: s2 (level: default)] <ts: 1712149259>
2024-04-03T13:01:00Z D [rs/rs101:27017] [backup/2024-04-03T13:00:59Z] init backup meta
2024-04-03T13:01:00Z I [rs/rs103:29017] got epoch {1712148849 3}
2024-04-03T13:01:00Z D [rs/rs101:27017] [backup/2024-04-03T13:00:59Z] nomination list for rs: [[rs102:28017 rs103:29017] [rs101:27017]]
2024-04-03T13:01:00Z D [rs/rs101:27017] [backup/2024-04-03T13:00:59Z] nomination rs, set candidates [rs102:28017 rs103:29017]
2024-04-03T13:01:00Z I [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] backup started
2024-04-03T13:01:00Z D [rs/rs103:29017] [backup/2024-04-03T13:00:59Z] skip after nomination, probably started by another node
2024-04-03T13:01:01Z D [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] flush incremental backup history
2024-04-03T13:01:01Z D [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] backup cursor id: 45d13b6e-84a5-448c-a7bf-1caca7035795
2024-04-03T13:01:04Z D [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] set journal up to {1712149261 2}
2024-04-03T13:01:05Z I [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] uploading data
2024-04-03T13:01:05Z D [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] stop cursor polling: <nil>, cursor err: <nil>
2024-04-03T13:01:05Z I [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] mark RS as error `upload data files: upload file `/data/db/journal/WiredTigerLog.0000000006`: get file stat: stat /data/db/journal/WiredTigerLog.0000000006: no such file or directory`: <nil>
2024-04-03T13:01:05Z I [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] mark backup as error `upload data files: upload file `/data/db/journal/WiredTigerLog.0000000006`: get file stat: stat /data/db/journal/WiredTigerLog.0000000006: no such file or directory`: <nil>
2024-04-03T13:01:05Z E [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] backup: upload data files: upload file `/data/db/journal/WiredTigerLog.0000000006`: get file stat: stat /data/db/journal/WiredTigerLog.0000000006: no such file or directory
2024-04-03T13:01:05Z D [rs/rs102:28017] [backup/2024-04-03T13:00:59Z] releasing lock
2024-04-03T13:01:05Z D [rs/rs101:27017] [backup/2024-04-03T13:00:59Z] bcp nomination: rs won by rs102:28017
2024-04-03T13:01:05Z D [rs/rs101:27017] [backup/2024-04-03T13:00:59Z] skip after nomination, probably started by another node

Specifically, I’m unable to generate incremental backups. While I can locate the file /data/db/journal/WiredTigerLog.0000000006 on one of the MongoDB nodes (rs103), I can’t seem to find it on the other nodes.

Here is the actual PBM status:

$ pbm status
Cluster:
========
rs:
  - rs/rs101:27017 [P]: pbm-agent v2.4.1 OK
  - rs/rs102:28017 [S]: pbm-agent v2.4.1 OK
  - rs/rs103:29017 [S]: pbm-agent v2.4.1 OK

PITR incremental backup:
========================
Status [ON]

Currently running:
==================
(none)

and the backup config:

storage:
  type: s3
  s3:
    endpointUrl: "http://xxxxxxxxxx"
    region: us-east-1
    bucket: mongo-backup
    credentials:
      access-key-id: xxxxxxx
      secret-access-key: xxxxxxx

I’m wondering if anyone could shed some light on what might be going wrong or if there’s something I might have missed in the setup process.

Thank you in advance for your help and suggestions!

3

Hi, for physical backups (required for incremental) the pbm agent container needs to be able to access the files of the mongod container. I think you will need to use either a bind mount or shared volume for storage.