Messages in this thread | | | Date | Wed, 26 Jan 2000 11:30:38 +1300 | From | Michael Clark <> | Subject | Re: [NFS] New version of Linux NFSv3 client is out... |
| |
Alan Cox wrote:
> It is up to the client to avoid bypassed writes on its own write cache, > inter client coherency isnt an NFS server issue. Basically NFS doesn't > even try to get it right. Clients tend to use lockf as barrier points to > commit. (ie flush before lockf, commit before unlock) but don't bet on it
So getting back to the original benchmark; I guess it would benefit from a lock/unlock on the dataset around each of the tests to avoid each test skewing the results of the next ie. the client should drop its cache with the lock at the beginning of the seq block test and force it to re-read from the server ?? (if not, i think this should be the desired behaviour).
I don't currently have a test setup with a NFSv3 server so i can't try this right now but cw and I will soon be performing some testing against NetApp Filers (with hopefully more reliable locking than Slowlaris).
Can someone point me at the source for the latest nfs-tests.
cheers, ~mc
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |