US 12,321,368 B2
Method, device, and storage medium for scheduling a distributed database
Gang Zhao, Beijing (CN); Qiang Chen, Beijing (CN); Ziqi Gao, Beijing (CN); and Kang Li, Beijing (CN)
Assigned to DOUYIN VISION CO., LTD., Beijing (CN)
Filed by Douyin Vision Co., Ltd., Beijing (CN)
Filed on Jun. 14, 2024, as Appl. No. 18/744,410.
Claims priority of application No. 202310928308.1 (CN), filed on Jul. 26, 2023.
Prior Publication US 2025/0036653 A1, Jan. 30, 2025
Int. Cl. G06F 16/27 (2019.01); G06F 16/21 (2019.01)
CPC G06F 16/278 (2019.01) [G06F 16/214 (2019.01)] 18 Claims
OG exemplary drawing
 
1. A method of scheduling a distributed database, comprising:
detecting load states of respective storage units in a first storage node in a distributed database, and determining a hotspot partition of the first storage unit based on the load states;
generating a replica splitting task for a target replica stored in the hotspot partition, the replica splitting task comprising a target number of subreplicas of the target replica after splitting and migration target addresses of the subreplicas; and
executing the replica splitting task by invoking a thread in a thread pool, to split the target replica into the subreplicas based on the target number of the subreplicas and migrate at least one of the subreplicas based on the migration target addresses of the subreplicas,
wherein generating the replica splitting task for the target replica stored in the hotspot partition comprises:
determining the target number of the subreplicas and the migration target addresses of the subreplicas based on a non-hotspot partition of the first storage unit and/or a further storage unit in the first storage node, wherein the migration target addresses of the subreplicas comprise one or more non-hotspot partitions of the first storage unit and/or the further storage unit in the first storage node.