diff options
| author | pjd <pjd@FreeBSD.org> | 2004-08-21 18:11:46 +0000 |
|---|---|---|
| committer | pjd <pjd@FreeBSD.org> | 2004-08-21 18:11:46 +0000 |
| commit | c3c6740d1ae2ca36dd507650dbed6962cc1b206d (patch) | |
| tree | c7a13feeb4a36d3ad1e0855d38be72d1ee949ab6 /sbin/geom/class/raid3/graid3.8 | |
| parent | fbfced34e485ed42ca65816dbf7029b29af13d53 (diff) | |
| download | FreeBSD-src-c3c6740d1ae2ca36dd507650dbed6962cc1b206d.zip FreeBSD-src-c3c6740d1ae2ca36dd507650dbed6962cc1b206d.tar.gz | |
Implement new reading algorithm, which will use parity component for reading
as well, even if device is in complete state.
I observe 40% of speed-up with this option for random read operations,
but slowdown for sequential reads.
Basically, without this option reading from a RAID3 device built from 5
components (c0-c4) looks like this:
Request no. Used components
1 c0+c1+c2+c3
2 c0+c1+c2+c3
3 c0+c1+c2+c3
With the new feature:
Request no. Used components
1 c0+c1+c2+c3
2 (c1^c2^c3^c4)+c1+c2+c3
3 c0+(c0^c2^c3^c4)+c2+c3
4 c0+c1+(c0^c1^c3^c4)+c3
5 c0+c1+c2+(c0^c1^c2^c4)
6 c0+c1+c2+c3
[...]
Diffstat (limited to 'sbin/geom/class/raid3/graid3.8')
| -rw-r--r-- | sbin/geom/class/raid3/graid3.8 | 22 |
1 files changed, 16 insertions, 6 deletions
diff --git a/sbin/geom/class/raid3/graid3.8 b/sbin/geom/class/raid3/graid3.8 index f38732d..713f979 100644 --- a/sbin/geom/class/raid3/graid3.8 +++ b/sbin/geom/class/raid3/graid3.8 @@ -24,7 +24,7 @@ .\" .\" $FreeBSD$ .\" -.Dd Aug 18, 2004 +.Dd Aug 20, 2004 .Dt GRAID3 8 .Os .Sh NAME @@ -33,7 +33,7 @@ .Sh SYNOPSIS .Nm .Cm label -.Op Fl hnv +.Op Fl hnrv .Ar name .Ar prov .Ar prov @@ -46,7 +46,7 @@ .Op Ar prov Op Ar ... .Nm .Cm configure -.Op Fl adhnv +.Op Fl adhnrRv .Ar name .Nm .Cm rebuild @@ -106,6 +106,12 @@ Additional options include: Hardcode providers' names in metadata. .It Fl n Turn off autosynchronization of stale components. +.It Fl r +Use parity component for reading in round-robin fashion. +Without this option parity component is not used at all for reading operations +when device is in complete state. +With this option specified random I/O read operations are even 40% faster, +but sequential reads are slower. .El .It Cm clear Clear metadata on the given providers. @@ -122,6 +128,10 @@ Do not hardcode providers' names in metadata. Hardcode providers' names in metadata. .It Fl n Turn off autosynchronization of stale components. +.It Fl r +Turn on round-robin reading. +.It Fl R +Turn off round-robin reading. .El .It Cm rebuild Rebuild the given component forcibly. @@ -167,10 +177,10 @@ Additional options include: Be more verbose. .El .Sh EXAMPLES -Use 3 disks to setup a RAID3 array. Create a file system, mount it, then unmount -it and stop device: +Use 3 disks to setup a RAID3 array (use round-robin reading feature). +Create a file system, mount it, then unmount it and stop device: .Bd -literal -offset indent -graid3 label -v data da0 da1 da2 +graid3 label -v -r data da0 da1 da2 newfs /dev/raid3/data mount /dev/raid3/data /mnt [...] |
