• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 859
  • Last Modified:

Linux NFS share not mounting from vSphere 4.1 host

When trying to mount a NFS Linux  share from the vSphere 4.1 host I get the below.

NFS Errors 149, 946, 160

The same NFS shares are accessible from the another host.

I can ping the NFS server from the host. Netcat reaches the NFS server from the host.

The NFS server can also ping the VMware host and resolve DNS.


Dec 22 10:07:19 vmkernel: 1143:21:50:57.036 cpu2:4310573)NFS: 149: Command: (mount) Server: (192.168.20.10) IP: (192.168.30.10) Path: (/finkdr) Label: (backup) Options: (None)
Dec 22 10:07:49 vmkernel: 1143:21:51:27.089 cpu3:4310573)WARNING: NFS: 946: MOUNT RPC failed with RPC status 13 (RPC was aborted due to timeout) trying to mount Server (192.168.20.10) Path (/finkdr)
Dec 22 10:07:49 vobd: Dec 22 10:07:49.956: 98833888868410us: [esx.problem.vmfs.nfs.mount.connect.failed] Failed to mount to server 192.168.20.10 mount point /finkdr. Error: Unable to connect to NFS server.
Dec 22 10:07:49 vmkernel: 1143:21:51:27.089 cpu3:4310573)NFS: 160: NFS mount 192.168.30.10:/finkdr failed: Unable to connect to NFS server
0
lhrslsshahi
Asked:
lhrslsshahi
2 Solutions
 
Patrick BogersDatacenter platform engineer LindowsCommented:
This issue may occur if DNS is unavailable and the RPC request times out during the DNS lookup.

 Have a quick read here for possible solutions.
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
can you vmkping the Linux NFS server/workstation?

do you also have a VMKernel defined, which is used to carry the traffic?

also see these tips

VMware KB: Unable to mount NFS datastore

VMware KB: Adding an NFS datastore to an ESX/ESXi host fails with error

VMware KB: Cannot connect to NFS network share
0
 
lhrslsshahiAuthor Commented:
I have taken the entries and from the /etc/resolv.conf on the host and NFS server
and have placed in the /etc/hosts still no joy.

I can vmkping NFS by ip and name and is successful.

The traffic passes on the management vmk0 same as the other working host

It was working last week until I had to rebuild the VM no changes have been made on the host and now only is accessible from one host!

As a test I have created a Windows NFS share same problem can be accessed from the other host
0
Transaction-level recovery for Oracle database

Veeam Explore for Oracle delivers low RTOs and RPOs with agentless transaction log backup and transaction-level recovery of Oracle databases. You can restore the database to a precise point in time, even to a specific transaction.

 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
okay....check these videos and see if you have missed anything..there is no audio..

these are my videos creating especially to show, how to setup, and what happens if you have no VMKernel Interface.

Video Tutorial - Creating an NFS Datastore using Windows 2008 R2 for VMware vSphere ESX/ESX 4.x

Video Tutorial - Adding an NFS Datastore to a VMware vSphere ESX/ESXi 4.x host server
0
 
lhrslsshahiAuthor Commented:
Andy,

Thanks for these videos I went through the second video, you created a seperate
VMKernel Interface, I tried that it didn't make any difference.  We have about 15 hosts that have no dedicated VMKernel Interface and have had no issues with Windows or Linux NFS shares before.  It's always been under the VMKernel Interface - Management

About to hit the wall! :-)
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
As long as your VMKernel Interface is on the same network as your NFS NAS.

which if you can vmkping the NAS, that indicates traffic can reach the NFS share.

Jumbo Frames are not enabled when they should be enabled?

You stated it worked until you re-built the VM - this VM is the NFS server?

So what is the difference between this working single hosts and all the others?

IP Address (if using) for authentication, is enabled?

the same share name is being used?

Can you upload screenshots of networking on working and non-working hosts?

just looking at these IP Addresses

192.168.20.10 - this is the server?

192.168.30.10 - this is the NAS (NFS)

different networks? routing and subnet masks are correct?
0
 
lhrslsshahiAuthor Commented:
The VM 192.168.30.10 is the NFS server (192.168.20.10 is a typo)

The problem host is 192.168.31.2 and the working host 192.168.30.5

The MTU is set to 1500 for switches on both hosts (Jumbo frames not enabled)

Same share name as before.

I have attached screenshots.
problem-host.PNG
workinghost.PNG
0
 
Paul SolovyovskyCommented:
keep in mind no root squash needs to be enabled so that there is no authentication occuring during connection
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
working host is:-

192.168.30.5

non-working is

192.168.31.2

Linux NFS is 192.168.30.10

so if you put your host as 192.168.30.11 does it work?

why the IP Address change to 192.168.31.2 ?
0
 
lhrslsshahiAuthor Commented:
Never had to use no_root )squash before, have tried it didn't work.

Unfortunately can't change the ip host as its in Production.

The VMs that reside on the 192.168.31.2 are on a different domain, it's been working for 4 years not entirely sure why its stopped working other than me rebuilding the NFS Server!
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
add a new VMKernel Interface as a test on 192.168.30.11

should not affect production?

so the IP Addresses for your hosts have always been...

192.168.30.5 and 192.168.31.2 ?

no other network changes or firewall changes have taken place on the network, at the same time the NFS server was rebuilt?
0
 
lhrslsshahiAuthor Commented:
OMG!! That worked addding VMKernel Interface as 192.168.30.11.

Thanks Andy, life saver, not sure how it was working before!?
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Maybe you had another VMKernel Interface set?

traffic is not being routed correctly....from 192.168.30.x to 192.168.31.x, subnet mask incorrect or something.
0
 
lhrslsshahiAuthor Commented:
I honestly don't remember having  a VMKernel Interface.  Oh well.. Just want to thank you for your help. Have a good Christmas and new year.
0
 
lhrslsshahiAuthor Commented:
Andy is always on the money
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
no problems, Merry Chrimbo.
0

Featured Post

VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now