Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.
Original topic: DM 全量导出数据的时候会把导出的数据放在/tidb-deploy/dm-worker-8262/ 里面
[TiDB Usage Environment] Production Environment / Testing / PoC
[TiDB Version]
[Encountered Issues]
[Reproduction Path] What operations were performed that led to the issue
[Issue Phenomenon and Impact]
[Attachment]
May I ask, when DM exports data in full, it places the exported data in /tidb-deploy/dm-worker-8262/. This is somewhat unreasonable because the file system space of /tidb-deploy/ itself is relatively small. Is there currently any parameter that can control the storage path of the full exported data?
The dir
option for the loader specifies the directory where the dump processing unit outputs SQL files, and it is also the directory where the load processing unit reads files. The default value for this configuration item is “./dumped_data”. Different tasks corresponding to the same instance must be configured with different directories. The path for dir
can be an absolute path. When using a relative path, a dir_name.task_name
directory will be generated in the deploy directory of each worker.
Okay, I have already checked this on the official website. I would like to ask, if DM reads MySQL’s binlog to synchronize incremental data, will DM write the binlog read from MySQL to disk, or will it directly apply the binlog read from MySQL to the downstream TiDB?
Enabling DM’s relay log will write to disk.
If DM relay log is not enabled, it should not be written to disk, right?
Correct. However, it will also write some temporary files, such as the table schema (incremental migration). It is recommended to reserve some disk space.
How large is the data volume? It’s best to reserve some space.