I just setup a ReadNAS 102 as an NFS file store for a VMware ESXi 4.0 server. I have both the server and NAS on a gig network switch and VMware able to read/write to the NAS device. I've tried several times to copy a 50GB VMware directory to the NAS device via both vSphere client as well as from command line and the NAS device eventually disconnects from NFS. I can still ping and access the NAS device but NFS is not responding. This has happened around the 25GB mark every time I've tried. I also noticed that the copy slowed down significantly the longer it went on. I've actually had to reboot the NAS device in order to get it mounted successfully to VMware again.
Today I quickly setup a Linux box as another NFS server to VMware. I copied the same VM directories to the Linux box quickly and without any problems. I think there must be something wrong with the ReadyNAS NFS implementation. Do I make NFS changes on the NAS device to make this work without eventually dyeing?
I think this is normal that the speed decreases over time, since most softwares actually show some kind of average. The non-responding thing makes me think of what I had before doing the factory reset for the 6.1.5 version that installed the latest BTRFS format. While copying the NAS just shut down, no interface, no shares available but ping still worked most of the time. The only way to get back up and running quickly was to power cycle.
I also use NFS for VMWare, but only for reading though. I run ESX 5.5 and found no problem while reading ISO. But they are smaller than 25GB, so I tried to use it as a writeable storage to give you some feedback and hit a copy problem too. It is less problematic than yours though. VMWare copied 45GB then stopped without error on a single 84GB VMDK file (so roughly half the data copied like for you), just like if I closed the window... My NAS is on heavy load right now and I will wait for it to finish its work before trying again. I'm refreshing my backups so this may take a while before I can try again, but if you got nothing by Monday evening this would mean I forgot, so feel free to ask for it in this case. The file was copied using a windows VM that runs the client (not via Vcenter web interface) and I copied from the local disk datastore (copy / paste from a datastore to another) to the NFS datastore.
I'm running 6.2.0 first beta, it may have some importance though I really doubt it. I will update to the second beta, that the update software found while I was configuring the NFS share, later this week (I want to check all my backups before).
Models such as the 300 series are generally a better fit as a datastore for virtualisation.
I don't think this is likely an issue with NFS. How many threads did you enable for NFS?
Did you enable snapshots for the shares, checksums etc.? If so, can you create a new share with those turned off and use that and see if it is any different?
Tried again after updating to 6.2.0-T1336, same result, copy stopped after 45Gb/84Gb. Progress bar just rushed to the end with no error (like if the transfer speed was multiplied by 100), so I'm thinking this is more a VMWare problem than a readyNAS problem. It would be easy to test on a linux, but I honestly don't have the will to do it as this is something I don't use for myself.
