From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; .NET CLR 1.0.3705; .NET CLR 1.1.4322) Description of problem: I have an nfs mounted a partition on a SnapAppliance SnapServer 4100 (I think its FreeBSD under the covers). I get I/O errors when copying files from the SnapServer to the Linx 9 installation. I also sometimes get data corruption. Repeating the same copy multiple times will work sometimes and fail others. In addition, a Thiz Linuz 7.0 installation I have works just fine with the same SnapAppliance SnapServer. Version-Release number of selected component (if applicable): How reproducible: Sometimes Steps to Reproduce: 1. mount snap server partition (mount 192.168.168.101:/root /snap1) 2. cp -r from to Actual Results: Get I/O error on some copied files. Also some files are corrupted on the destination machine. Expected Results: Expect not to get errors Additional info: Copying the same file twice does not produce the same output.
What transport are you using UDP or TCP? Are you using Soft mounts?
I'm using UDP. I've tried both soft and hard mounts. I tried TCP as well but got "mount: RPC: program not registered" from the mount command.
Would it be possible to get a ethereal trace of the problem?
Not really. I decided to try installing Red Had Enterprise ES and just discovered that that resolved the problem.
I had a similar problem with 2.4.20-27.7 and 2.4.20-24.7 but not 2.4.18-3smp (different machine), all RH7.3, connected to a SnapServer 1100. After I added the mount option -o nfsvers=2, the problem seems to have disappeared.
The only way to decisively tell why nfsv3 breaks a SnapServer is to have get a network trace. I need to see which NFS op is failing