Buddy allocator collets some statistics (ext4_mb_collect_stats()).
There are counters in ext4 superblock info:
- s_bal_reqs – number of requests > 1
- s_bal_allocated – found enough chunk
- s_bal_ex_scanned – how much extents scanned
- s_bal_goals – block is allocated from goal
- s_bal_breaks – allocator was interrupted (because of s_mb_max_to_scan)
- s_bal_2orders – 2^orders hits
Statistic is show on fs unmount (ext4_mb_release()) if /sys/fs/ext4/_partition_/mb_stats is set.
It looks that for the same IO rate s_bal_allocated/s_bal_ex_scanned can show how effective allocator at some period of time.
There is seq-file with details
[root@localhost cray-lustre]# cat /proc/fs/ldiskfs/loop1/mb_groups
#group: bfree gfree frags first pa [ 2^0 2^1 2^2 2^3 2^4 2^5 2^6 2^7 2^8 2^9 2^10 2^11 2^12 2^13 ]
#0 : 25017 25017 1 7367 0 [ 1 0 0 1 1 1 0 1 1 2 1 1 1 2 ]
#1 : 30397 31266 2 2048 2 [ 1 0 1 1 1 1 2 2 1 0 1 2 2 2 ]
#2 : 28352 28352 1 4096 0 [ 0 0 0 0 0 0 1 1 0 1 1 1 2 2 ]
#3 : 1632 1632 1 1024 0 [ 0 0 0 0 0 1 1 0 0 1 1 0 0 0 ]
English
Русский