Whamcloud - gitweb
LU-14641 osd-ldiskfs: write commit declaring improvement
authorWang Shilong <wshilong@ddn.com>
Mon, 26 Apr 2021 03:23:26 +0000 (11:23 +0800)
committerAndreas Dilger <adilger@whamcloud.com>
Wed, 21 Jul 2021 05:56:45 +0000 (05:56 +0000)
commitee5d5674a61d31675a8314e635214694671fdc75
treed35195d81ecfcee7a27b3b7e4203b68c0a558749
parent67f57106f5d2dd464d89ca7e5f8920e9939a0b4f
LU-14641 osd-ldiskfs: write commit declaring improvement

This patch try to:

1)extent bytes could be missed to increase with less than
1M, fix to to compare it with current value, and decay
it for every allocation.

2)with system space usage growing up, mballoc codes won't
try best to scan block group to align best free extent as
we can. So extent bytes per extent could be decayed to a
very small value, this could make us reserve too many credits.
We could be more optimistic in the credit reservations, even
in a case where the filesystem is nearly full, it is extremely
unlikely that the worst case would ever be hit.

3)Add extent bytes stats and debug ability to analysis
over reservation problem.

Lustre-commit: 0f81c5ae973bf7fba45b6ba7f9c5f4fb1f6eadcb
Lustre-change: https://review.whamcloud.com/43446

Signed-off-by: Wang Shilong <wshilong@ddn.com>
Change-Id: I357c4a855147ba26a9e9bbe9ab1269bcfd44e5f3
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/44244
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
lustre/osd-ldiskfs/osd_io.c
lustre/osd-ldiskfs/osd_lproc.c