Controls the number of AI model metadata entries that PolarDB caches in an LRU (Least Recently Used) cache, reducing reads to the model metadata table.
Data type
integer
Values
Default: 4
Minimum: 0
Maximum: 100
Description
By default, PolarDB caches metadata for up to 4 models. Increase this value when your workload accesses more models concurrently to reduce repeated reads to the model metadata table.
Important
A change to polar.ai.lru_entity_size takes effect only after you reconnect to the database.
Examples
Set the cache size for the current session:
SET polar.ai.lru_entity_size = 8;Set the cache size for all new connections to a specific database:
ALTER DATABASE database_name SET polar.ai.lru_entity_size = 10;