Is it necessary to perform a checksum for TiDB's BR backup?

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: tidb的br备份有必要去checksum吗?

| username: 舞动梦灵

Is it necessary to perform a checksum for TiDB’s BR backup? In the test environment, a 7GB backup took 6 minutes, and the checksum took 9 minutes. If the actual production backup is in the terabyte range, will the checksum take a very long time? Is it necessary?

| username: xfworld | Original post link

Try using version 6.x or 7.x, the speed of BR will be faster.

Checksum is to ensure data integrity and consistency… It is necessary…

| username: 像风一样的男子 | Original post link

Yes, my backup checksum is stuck for several hours now, but if I turn it off, I can’t guarantee the backup data is usable.

| username: 舞动梦灵 | Original post link

The BR version must be consistent with the cluster version, right? I checked the official documentation. A 4.0 version backup is for a 4.0 cluster. You can use the 5.0 version of BR to restore to a 5.0 cluster. The data is 15T… If the backup takes 24 hours, then the check will take more than a day, right? That’s too long.

| username: 舞动梦灵 | Original post link

So it’s really a dilemma. Migrating in production. Using BR backup, there is 15TB of data. If the backup takes a day, and then it takes another day and a half to verify, can we still compare the data from 24 hours ago at the backup time point during the verification?

| username: 像风一样的男子 | Original post link

If you are doing a migration, temporarily extend the GC time.

| username: 舞动梦灵 | Original post link

You mean this:
UPDATE mysql.tidb SET VARIABLE_VALUE = ‘720h’ WHERE VARIABLE_NAME = ‘tikv_gc_life_time’;

| username: xfworld | Original post link

Yes, the versions must match, otherwise, it cannot be restored.

| username: zhanggame1 | Original post link

There is no need to check the backup set. Just import it directly and then use the checking tool to compare the two databases.

| username: 大飞哥online | Original post link

In order to ensure the availability of the backup result set, it is still necessary to check.
Unless you perform a recovery drill immediately after the backup, you can skip it. After recovery, you can also verify the data.

| username: Fly-bird | Original post link

It is necessary, after all, to verify data consistency.

| username: ajin0514 | Original post link

This is indeed a problem.

| username: system | Original post link

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.