Doris supports backing up the current data in the form of files to the remote storage system like S3 and HDFS. Afterwards, you can restore data from the remote storage system to any Doris cluster through the restore command. Through this function, Doris can support periodic snapshot backup of data. You can also use this function to migrate data between different clusters.
This feature requires Doris version 0.8.2+
The backup operation is to upload the data of the specified table or partition directly to the remote warehouse for storage in the form of files stored by Doris. When a user submits a Backup request, the system will perform the following operations:
Snapshot and snapshot upload
The snapshot phase takes a snapshot of the specified table or partition data file. After that, backups are all operations on snapshots. After the snapshot, changes, imports, etc. to the table no longer affect the results of the backup. Snapshots only generate a hard link to the current data file, which takes very little time. After the snapshot is completed, the snapshot files will be uploaded one by one. Snapshot uploads are done concurrently by each Backend.
Metadata preparation and upload
After the data file snapshot upload is complete, Frontend will first write the corresponding metadata to a local file, and then upload the local metadata file to the remote warehouse through the broker. Completing the final backup job
Dynamic Partition Table Description
If the table is a dynamic partition table, the dynamic partition attribute will be automatically disabled after backup. When restoring, you need to manually enable the dynamic partition attribute of the table. The command is as follows:
ALTER TABLE tbl1 SET ("dynamic_partition.enable"="true")
colocate_with property of a table.Create a hdfs remote warehouse example_repo (S3 skips step 1):
CREATE REPOSITORY `example_repo` WITH HDFS ON LOCATION "hdfs://hadoop-name-node:54310/path/to/repo/" PROPERTIES ( "fs.defaultFS"="hdfs://hdfs_host:port", "hadoop.username" = "hadoop" );
Create a remote repository for S3 : s3_repo (HDFS skips step 2)
CREATE REPOSITORY `s3_repo`
WITH S3
ON LOCATION "s3://bucket_name/test"
PROPERTIES
(
"AWS_ENDPOINT" = "http://xxxx.xxxx.com",
"AWS_ACCESS_KEY" = "xxxx",
"AWS_SECRET_KEY" = "xxx",
"AWS_REGION" = "xxx"
);
Note that.
ON LOCATION is followed by Bucket Name here
Full backup of table example_tbl under example_db to warehouse example_repo:
BACKUP SNAPSHOT example_db.snapshot_label1 TO example_repo ON (example_tbl) PROPERTIES ("type" = "full");
Under the full backup example_db, the p1, p2 partitions of the table example_tbl, and the table example_tbl2 to the warehouse example_repo:
BACKUP SNAPSHOT example_db.snapshot_label2 TO example_repo ON ( example_tbl PARTITION (p1,p2), example_tbl2 );
View the execution of the most recent backup job:
mysql> show BACKUP\G; *************************** 1. row *************************** JobId: 17891847 SnapshotName: snapshot_label1 DbName: example_db State: FINISHED BackupObjs: [default_cluster:example_db.example_tbl] CreateTime: 2022-04-08 15:52:29 SnapshotFinishedTime: 2022-04-08 15:52:32 UploadFinishedTime: 2022-04-08 15:52:38 FinishedTime: 2022-04-08 15:52:44 UnfinishedTasks: Progress: TaskErrMsg: Status: [OK] Timeout: 86400 1 row in set (0.01 sec)
View existing backups in remote repositories:
mysql> SHOW SNAPSHOT ON example_repo WHERE SNAPSHOT = "snapshot_label1"; +-----------------+---------------------+--------+ | Snapshot | Timestamp | Status | +-----------------+---------------------+--------+ | snapshot_label1 | 2022-04-08-15-52-29 | OK | +-----------------+---------------------+--------+ 1 row in set (0.15 sec)
For the detailed usage of BACKUP, please refer to here.
Currently, we support full backup with the smallest partition (Partition) granularity (incremental backup may be supported in future versions). If you need to back up data regularly, you first need to plan the partitioning and bucketing of the table reasonably when building the table, such as partitioning by time. Then, in the subsequent running process, regular data backups are performed according to the partition granularity.
Users can back up the data to the remote warehouse first, and then restore the data to another cluster through the remote warehouse to complete the data migration. Because data backup is done in the form of snapshots, new imported data after the snapshot phase of the backup job will not be backed up. Therefore, after the snapshot is completed and until the recovery job is completed, the data imported on the original cluster needs to be imported again on the new cluster.
It is recommended to import the new and old clusters in parallel for a period of time after the migration is complete. After verifying the correctness of data and services, migrate services to a new cluster.
SHOW PARTITIONS FROM table_name; and SHOW TABLETS FROM table_name; to view the number of shards in each partition and the number of file versions in each shard to estimate job execution time. The number of files has a great impact on the execution time of the job. Therefore, it is recommended to plan partitions and buckets reasonably when creating tables to avoid excessive sharding.SHOW BACKUP or SHOW RESTORE command. It is possible to see error messages in the TaskErrMsg column. But as long as the State column is not CANCELLED, the job is still continuing. These tasks may retry successfully. Of course, some Task errors will also directly cause the job to fail.COMMIT phase of the recovery job, the overwritten data on the current cluster may no longer be restored. If the restore job fails or is canceled at this time, the previous data may be damaged and inaccessible. In this case, the only way to do it is to perform the recovery operation again and wait for the job to complete. Therefore, we recommend that if unnecessary, try not to restore data by overwriting unless it is confirmed that the current data is no longer used.The commands related to the backup and restore function are as follows. For the following commands, you can use help cmd; to view detailed help after connecting to Doris through mysql-client.
CREATE REPOSITORY
Create a remote repository path for backup or restore. Please refer to Create Repository Reference.
BACKUP
Perform a backup operation. Please refer to Backup Reference.
SHOW BACKUP
View the execution of the most recent backup job. Please refer to Show Backup Reference。
SHOW SNAPSHOT
View existing backups in the remote repository. Please refer to Show Snapshot Reference.
CANCEL BACKUP
Cancel the currently executing backup job. Please refer to [Cancel Backup Reference] (../../sql-manual/sql-statements/Data-Definition-Statements/Backup-and-Restore/CANCEL-BACKUP.md).
DROP REPOSITORY
Delete the created remote repository. Deleting a warehouse only deletes the mapping of the warehouse in Doris, and does not delete the actual warehouse data. Please refer to [Drop Repository Reference] (../../sql-manual/sql-statements/Data-Definition-Statements/Backup-and-Restore/DROP-REPOSITORY.md).
For more detailed syntax and best practices used by BACKUP, please refer to the BACKUP command manual, You can also type HELP BACKUP on the MySql client command line for more help.