Adding NFS storage fails

[markok]

I have NFS storage waiting and when I try to mount NFS storage with VI Client, I will get an error message:

"Error during the configuration of the host: Unable to get FS Attrs for /vmfs/volumes/aeae6580-0c7aa095"

Host is VMware ESX Server v3.0.1 with latest patches (VC is also with latest patches) and NFS storage is from NetApp. I found similar problem, but no answer: http://www.vmware.com/community/thread.jspa?messageID=633853򚯽

The VMKernel interface (for NFS and iSCSI) is configured. Also some SAN related KB articles (http://kb.vmware.com/kb/6482648) but no help from there.

Here is some messages from the ESX Server hostd.log:

\----

---

'FSVolumeProvider' 5282736 verbose] AddNasVolume called

'FSVolumeProvider' 5282736 verbose] LookupVmfs: Cannot find VMFS volume with id , or name NfsStorage

'FSVolumeProvider' 5282736 verbose] LookupNasByName: Cannot find NasVolume with volume name NfsStorage

'FSVolumeProvider' 5282736 verbose] VmkCtl mount or ProcessNas failed: Unable to get FS Attrs for /vmfs/volumes/aeae6580-0c7aa095

\----

---

Any ideas what is wrong?

[Texiwill]

Hello,

It appears that your NetAPP NFS server is unreachable, or not responding properly. Do you have a vmkernel interface for NFS/iSCSI/ and vMotion? If you do then depending on the IP and netmask of the interface you may be going out the wrong pipe. For example my setup has the following:

vMotion 10.0.3.0/24 network

iSCSI 10.0.2.0/24 network

NAS 10.0.1.0/24 network

The Service COnsole has a 10.0.0.0/16 network. The reason for this is that the SC and iSCSI must be able to see each other. Half the iSCSI code resides on the service console.

If you do not have the proper network masks and gateways then you may be going out the wrong pipe. I had my vMotion set as a /16 as well and went out that pipe until I fixed it. So double check your network/masks for each of the 3 vmkernel networks.

Also, look at the logs on your netapp. Is it seeing the server try to connect? Your permissions on the netapp may have to be by server FQDN, server name, or even IP. If you do not see an attempt to connect then I would say it is a netmask error on the ESX server. If you do see it attempt to connect then it could be a permission problem on the netapp server.

Also, is netapp speaking NFS v2? or v3? ESX supports v2 over UDP. If you are trying to use TCP instead it may not work.

In addition, the delegate user inside ESX is 'root' or UID 0. Be sure you are allowing this user to read/write the store. You may need to assign a new delegate user instead. Page 232 of http://www.vmware.com/pdf/vi3_301_201_server_config.pdf.

Best regards,

Edward

[markok]

Hi Edward and thanks for your reply.

Do you really mean NFSv2 over UDP? The manual (same manual you pointed) says in page 133 (Configuring Storage: Network Attached Storage) that ESX supports only NFSv3 over TCP[/u].

I am aware of iSCSI and SC dependency and I have already successfully used software iSCSI and now I want to try NFS. One ESX has already configured with iSCSI against NetApp filer and now I want second ESX to talk NFS with NetApp.

NetApp shows that connection attempts are made, so there shouldn't be any network issues there. I have SC, VM traffic and VMkernel for NFS connected to one physical NIC. I know this is kind of weird situation, but I just wanted quick and dirty setup to verify NFS functionality. VMotion network is in separated network, but all other goes thru one common network.

Almost all possible settings have been tried already at NetApp side, but without any luck. Also the NFS share was allowed for all users with root permissions, but that didn't help either. The error message comes so quickly that it makes me wonder if it's in ESX side.

Both ESX side and NetApp side configurations doesn't have so much options, so it should be quite straight forward situtation. But it doesn't seem to be so straight after all.

[Texiwill]

Hello,

My NFS server speaks NFS v3 UDP and TCP. You know I am not sure I can tell which is actually in use... It is software and very limited in that scope, but enable UDP and TCP on the netapp, it will not hurt.

If your NetApp shows the connections, that is very good. What else does it say about the connections? Any errors at all? My ESX servers are constantly sending NFS3_GETATTR requests, be sure all requests are allowed. Does the netapp show which IP is being used for attachment? Does it jive with the IP you setup for the NFS vmkernel portgroup off your SC vSwitch? Or is it showing the SC IP?

I would increase verbosity on the netapp and also inspect the /var/log/vmkernel log file for issues. B

I think it is a networking issue, and most likely vSwitch related in some way.

Best regards,

Edward

[markok]

Hi

I have solved the problem. The problem was actually very stupid. The NetApp filer NFS share was full and ESX couldn't write anything into it.

It's funny that NetApp filer didn't warn about that in console or in GUI at all. It just created a new NFS share that was already full and didn't even hinted about it.

Thanks for helping me on this...

[markok]

Creating a NFS storage fails in Virtual Center and error message "Error during the configuration of the host: Unable to get FS Attrs for /vmfs/volumes/aeae6580-0c7aa095" was shown. Problem was that NetApp NFS share was full and the ESX server couldn't write into it.