From 0d12134b55a4f277ac37709e57714742be0e02f1 Mon Sep 17 00:00:00 2001 From: qiancai Date: Tue, 30 Sep 2025 15:30:56 +0800 Subject: [PATCH 1/2] Update br-snapshot-guide.md --- br/br-snapshot-guide.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/br/br-snapshot-guide.md b/br/br-snapshot-guide.md index 5924636b7708a..3ac88ebfd77e3 100644 --- a/br/br-snapshot-guide.md +++ b/br/br-snapshot-guide.md @@ -264,7 +264,7 @@ The impact of backup on cluster performance can be reduced by limiting the backu ### Performance and impact of snapshot restore - During data restore, TiDB tries to fully utilize the TiKV CPU, disk IO, and network bandwidth resources. Therefore, it is recommended to restore the backup data on an empty cluster to avoid affecting the running applications. -- The speed of restoring backup data is much related with the cluster configuration, deployment, and running applications. In internal tests, the restore speed of a single TiKV node can reach 100 MiB/s. The performance and impact of snapshot restore are varied in different user scenarios and should be tested in actual environments. +- The speed of restoring backup data is much related with the cluster configuration, deployment, and running applications. In internal tests, the restore speed of a single TiKV node can reach 1 GiB/s. The performance and impact of snapshot restore are varied in different user scenarios and should be tested in actual environments. - BR provides a coarse-grained Region scattering algorithm to accelerate Region restore in large-scale Region scenarios. This algorithm ensures that each TiKV node receives stable and evenly distributed download tasks, thus fully utilizing the resources of each TiKV node and achieving a rapid parallel recovery. In several real-world cases, the snapshot restore speed of the cluster is improved by about 3 times in large-scale Region scenarios. - Starting from v8.0.0, the `br` command-line tool introduces the `--tikv-max-restore-concurrency` parameter to control the maximum number of files that BR downloads and ingests per TiKV node. By configuring this parameter, you can also control the maximum length of the job queue (the maximum length of the job queue = 32 \* the number of TiKV nodes \* `--tikv-max-restore-concurrency`), thereby controlling the memory consumption of the BR node. From e727dea250a471db4387cb76b0e62c6651e330ad Mon Sep 17 00:00:00 2001 From: Grace Cai Date: Thu, 9 Oct 2025 13:39:52 +0800 Subject: [PATCH 2/2] Update br/br-snapshot-guide.md --- br/br-snapshot-guide.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/br/br-snapshot-guide.md b/br/br-snapshot-guide.md index 3ac88ebfd77e3..9ae1f265d797f 100644 --- a/br/br-snapshot-guide.md +++ b/br/br-snapshot-guide.md @@ -264,7 +264,7 @@ The impact of backup on cluster performance can be reduced by limiting the backu ### Performance and impact of snapshot restore - During data restore, TiDB tries to fully utilize the TiKV CPU, disk IO, and network bandwidth resources. Therefore, it is recommended to restore the backup data on an empty cluster to avoid affecting the running applications. -- The speed of restoring backup data is much related with the cluster configuration, deployment, and running applications. In internal tests, the restore speed of a single TiKV node can reach 1 GiB/s. The performance and impact of snapshot restore are varied in different user scenarios and should be tested in actual environments. +- The speed of restoring backup data is much related with the cluster configuration, deployment, and running applications. The performance and impact of snapshot restore are varied in different user scenarios and should be tested in actual environments. - BR provides a coarse-grained Region scattering algorithm to accelerate Region restore in large-scale Region scenarios. This algorithm ensures that each TiKV node receives stable and evenly distributed download tasks, thus fully utilizing the resources of each TiKV node and achieving a rapid parallel recovery. In several real-world cases, the snapshot restore speed of the cluster is improved by about 3 times in large-scale Region scenarios. - Starting from v8.0.0, the `br` command-line tool introduces the `--tikv-max-restore-concurrency` parameter to control the maximum number of files that BR downloads and ingests per TiKV node. By configuring this parameter, you can also control the maximum length of the job queue (the maximum length of the job queue = 32 \* the number of TiKV nodes \* `--tikv-max-restore-concurrency`), thereby controlling the memory consumption of the BR node.