MyFavMirrors/minio
1
0
Fork 0
mirror of https://github.com/minio/minio.git synced 2025-11-07 12:52:58 -05:00
Code Issues Packages Projects Releases Wiki Activity

improve performance for inlined data (#15603)

Browse Source 
inlined data often is bigger than the allowed
O_DIRECT alignment, so potentially we can write
'xl.meta' without O_DSYNC instead we can rely on
O_DIRECT + fdatasync() instead.

This PR allows O_DIRECT on inlined data that
would gain the benefits of performing O_DIRECT,
eventually performing an fdatasync() at the end.

Performance boost can be observed here for small
objects < 128KiB. The performance boost is mainly
seen on HDD, and marginal on NVMe setups.
This commit is contained in:
Harshavardhana
2022-08-29 11:19:29 -07:00
committed by GitHub
parent 92a0a59de2
commit 97376f6e8f
2 changed files with 28 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
internal/ioutil/odirect_reader.go
View File

@@ -40,9 +40,9 @@ type ODirectReader struct {
// Block sizes constant.
const (
BlockSizeSmall = 128 * humanize.KiByte // Default r/w block size for smaller objects.
BlockSizeLarge = 2 * humanize.MiByte // Default r/w block size for larger objects.
BlockSizeReallyLarge = 4 * humanize.MiByte // Default write block size for objects per shard >= 64MiB
BlockSizeSmall = 32 * humanize.KiByte // Default r/w block size for smaller objects.
BlockSizeLarge = 2 * humanize.MiByte // Default r/w block size for larger objects.
BlockSizeReallyLarge = 4 * humanize.MiByte // Default write block size for objects per shard >= 64MiB
)
// O_DIRECT aligned sync.Pool's