Enhancement

dt_bg_gc_check_interval and dt_bg_gc_max_segments_to_check_every_round is too conservative. It may cause too much outdate data.

For example, after we expand TiFlash cluster from 2 nodes to 10 nodes, the data on 2 old nodes are removed very slow by default settings. After changed those settings to the values below, the removal boost.

profiles.default.dt_bg_gc_check_interval : 5
profiles.default.dt_bg_gc_max_segments_to_check_every_round : 100