[Gluster-devel] large_read fuse option when direct-io mode is disabled?

Gordan Bobic gordan at bobich.net
Sat Jul 4 15:54:57 UTC 2009


Harshavardhana wrote:
> Prior to 2.6.26 fuse only supported single page write requests.  In theory all
> fuse filesystem should be able support bigger than 4k writes, as there's
> nothing in the API to prevent it. But there was a known case in ntfs-3g 
> which caused corruption while using earlier fuse so they disabled bigger 
> than 4kb writes by default since 2.6.26. But there is an option to use 
> it for filesystems which can handle it i.e -obig_writes as fuse opt args.
> 
>       So as earlier fuse 2.7.x and direct-io disabled glusterfs used to 
> receive 4kb chucks at all times resulting in applications writing at 
> bigger block sizes crunch with slower performance. But big_writes will 
> be enabled by default in glusterfs if you use fuse-2.8.0 even if you use 
> disable-direct-io or not.

So, this isn't applicable at all for kernels < 2.6.26?

Gordan





More information about the Gluster-devel mailing list