Disabling barriers helps with writes at the expense of loss of data security in the case of a power failure. For testing I'd say its definitely worth a play but probably not in a production environment. I don't have any numbers to hand on the differences unfortuantely - its just something I use on my test systems fairly generically these days. So far I've not had any bad experiences using it. On 04/05/11 14:07, Chris Webb wrote: > Haven <haven at thehavennet.org.uk> writes: > >> Running the same test here on the virtual I'm getting: >> 524288000 bytes (524 MB) copied, 15.2004 s, 34.5 MB/s >> >> Running that on the underlying drive of one of the cluster I get: >> 524288000 bytes (524 MB) copied, 7.54742 s, 69.5 MB/s > Yes, that's much more the sort of performance I was seeing in my previous > tests. I'm expecting some overhead from Sheepdog, but just not quite a > factor of ten! > >> I'm mounting the sheepdog partition using ext4 on an LVM partition using: >> noatime,barrier=0,user_xattr,data=writeback > I will re-test with ext4. Do you find disabling barriers is making a huge > difference? > > Cheers, > > Chris. |