Skip to content

pitr: add more description for performance parameters #21503

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

pitr: add more description for performance parameters #21503

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 7 additions & 0 deletions br/br-pitr-guide.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -159,6 +159,13 @@ The following steps describe how to clean up backup data that exceeds the backup
> External storage only contains KV data of a single replica. Therefore, the data size in external storage does not represent the actual data size restored in the cluster. BR restores all replicas according to the number of replicas configured for the cluster. The more replicas there are, the more data can be actually restored.
> The default replica number for all clusters in the test is 3.
> To improve the overall restore performance, you can modify the [`import.num-threads`](/tikv-configuration-file.md#import) item in the TiKV configuration file and the [`pitr-concurrency`](/br/use-br-command-line-tool.md#common-options) option in the BR command.
> When the upstream has **many Regions** and a **short flush interval**, PITR produces lots of small files. This can increase batching/dispatch overhead during restore. You can **moderately** increase the following to raise the number of files processed per batch and improve throughput:

- `pitr-batch-size`: cumulative **bytes per batch** (default **16 MiB**).
- `pitr-batch-count`: **number of files per batch** (default **8**).

> The two thresholds are **equivalent in effect** and **evaluated independently**: **whichever is reached first** closes the current batch and starts the next; once one is met, the **other is ignored** for that batch.
Comment on lines +162 to +167
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

low

Thanks for adding these details. For better clarity and to align with the style guide, I have a few suggestions for the wording:12

  • In the first sentence, "lots of" is a bit informal. "a large number of" would be better.
  • The last sentence is a bit complex and uses passive voice. It can be simplified for easier reading.

Style Guide References

Suggested change
> When the upstream has **many Regions** and a **short flush interval**, PITR produces lots of small files. This can increase batching/dispatch overhead during restore. You can **moderately** increase the following to raise the number of files processed per batch and improve throughput:
- `pitr-batch-size`: cumulative **bytes per batch** (default **16 MiB**).
- `pitr-batch-count`: **number of files per batch** (default **8**).
> The two thresholds are **equivalent in effect** and **evaluated independently**: **whichever is reached first** closes the current batch and starts the next; once one is met, the **other is ignored** for that batch.
> When the upstream has **many Regions** and a **short flush interval**, PITR produces a large number of small files. This can increase batching/dispatch overhead during restore. You can **moderately** increase the following to raise the number of files processed per batch and improve throughput:
- `pitr-batch-size`: cumulative **bytes per batch** (default **16 MiB**).
- `pitr-batch-count`: **number of files per batch** (default **8**).
> These two thresholds are evaluated independently. Whichever threshold is met first closes the current batch and starts the next. The other threshold is then ignored for that batch.

Footnotes

  1. Documentation should be clear, simple, and readable, with correct grammar. (link) ↩

  2. Avoid overusing passive voice. Prefer active voice for clarity. (link) ↩



Testing scenario 1 (on [TiDB Cloud](https://tidbcloud.com)) is as follows:

Expand Down