Bug #16689
closed
mkfs.xfs slow performance with discards and object map
Added by Jason Dillaman almost 8 years ago.
Updated over 7 years ago.
Description
Examples:
object-map enabled:
time mkfs.xfs -s size=4096 -f /dev/sda
real 9m10.882s
user 0m0.000s
sys 0m0.012s
object-map disabled:
real 0m1.780s
user 0m0.000s
sys 0m0.012s
Files
I am attaching 2 files for lttng trace run captured while running: mkfs.xfs -s size=4096 -f /dev/sda.
- replay.100G.slow.bin - trace with object-map enabled.
- replay.100G.fast.bin - trace with object-map disabled.
Other images features that were common across both runs: layering, striping, exclusive-lock, deep-flatten.
size: 102400 MB in 25600 objects
order: 22 (4096 kB objects)
format: 2
flags:
stripe unit: 4096 kB
stripe count: 1
Let me know if I can provide anything else. Thanks.
Attaching the raw lttng files for librbd runs. The new image is 10G instead of the 100G in the previous case, but all other attributes should match up.
- Status changed from New to In Progress
- Assignee set to Jason Dillaman
- Status changed from In Progress to Fix Under Review
- Status changed from Fix Under Review to Pending Backport
- Copied to Backport #16867: jewel: mkfs.xfs slow performance with discards and object map added
- Status changed from Pending Backport to Resolved
Also available in: Atom
PDF