Archive-Team
/
zfs
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases 103 Wiki Activity

Added information to zfs_vdev_aggregation_limit

Richard Elling 2019-02-13 09:49:28 -08:00
parent afefcbfed6
commit 133d411d95
1 changed files with 10 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
ZFS-on-Linux-Module-Parameters.md

@ -3131,6 +3131,15 @@ For writes, aggregation can occur at the ZFS or disk level.
`zfs_vdev_aggregation_limit` is the upper bound on the size of the larger, `zfs_vdev_aggregation_limit` is the upper bound on the size of the larger,
aggregated I/O. aggregated I/O.
Setting `zfs_vdev_aggregation_limit = 0` effectively disables aggregation by ZFS.
However, the block device scheduler can still merge (aggregate) I/Os. Also, many
devices, such as modern HDDs, contain schedulers that can aggregate I/Os.
In general, I/O aggregation can improve performance for devices, such as HDDs,
where ordering I/O operations for contiguous LBAs is a benefit. For random access
devices, such as SSDs, aggregation might not improve performance relative to the
CPU cycles needed to aggregate.
| zfs_vdev_aggregation_limit | Notes | zfs_vdev_aggregation_limit | Notes
|---|--- |---|---
| Tags | [vdev](#vdev), [ZIO_scheduler](#zio_scheduler) | Tags | [vdev](#vdev), [ZIO_scheduler](#zio_scheduler)
@ -3140,6 +3149,7 @@ aggregated I/O.
| Range | 0 to 131,072 (default) or 16,777,216 (if `zpool` `large_blocks` feature is enabled) | Range | 0 to 131,072 (default) or 16,777,216 (if `zpool` `large_blocks` feature is enabled)
| Default | 131,072 (128 KiB) | Default | 131,072 (128 KiB)
| Change | Dynamic | Change | Dynamic
| Verification | ZFS aggregation is observed with `zpool iostat -r` and the block scheduler merging is observed with `iostat -x`
| Versions Affected | all | Versions Affected | all
### zfs_vdev_cache_size ### zfs_vdev_cache_size