mirror of
https://github.com/versity/scoutfs.git
synced 2026-09-03 06:37:00 +00:00
The number of inodes that each inode group lock covers is a balance between minimizing lock overhead and contention. With 128 inodes per lock we're still doing quite a bit of bulk creates, finds, etc, per lock but could reduce the risk of contention a bit. Let's see how it works out. Signed-off-by: Zach Brown <zab@versity.com>