Invisible NFS mount problem

[aerius1]

Hello,

Seems like I have an old NFS mount on one of the ESXi hosts that are trying to access a NFS share that does not exist.

I do not want this share anymore. But I cannot find the share/mount in Vcenter and vSphere or when using the command "esxcli storage nfs list" when running this command it's empty.

But my vmkernel.log are filled with

2016-12-15T12:01:53.004Z cpu12:151307)WARNING: NFS: 221: Got error 13 from mount call

2016-12-15T12:02:23.007Z cpu15:221163)WARNING: NFS: 221: Got error 13 from mount call

Also getting the logs on the NAS that says the mountpoint doesn't exist. Which is OK because it doesn't.

Dec 15 13:09:23 <NAS-Name> mountd[22578]: mount request from <ESXi host IP> for non existent path <mountpoint>

Dec 15 13:09:23 <NAS-Name> mountd[22578]: mount request denied from <ESXi host IP> for <mountpoint>

What can I do to find this NFS mount? And remove it. So it stops trying to access the share every 30 seconds.

All help appreciated. If there is a thread and I have not found it. Please link me that because I have not found any.

Thanks,

Phil

[RAJ_RAJ]

H i,

Could you please try below

1. To list the permanently detached devices, run this command:
   
   # esxcli storage core device detached list
   
   You see output similar to:
   
   Device UID State
   ------------------------------------ -----
   naa.50060160c46036df50060160c46036df off
   naa.6006016094602800c8e3e1c5d3c8e011 off
   
   
2. To permanently remove the device configuration information from the system, run this command:
   
   # esxcli storage core device detached remove -d NAA_ID
   
   For example:
   
   # esxcli storage core device detached remove -d naa.50060160c46036df50060160c46036df

Reference # How to unmount a LUN or detach a datastore device from ESXi hosts (2004605) | VMware KB

[aerius1]

Hello,

Thank you for your answer. However I get no output on the command "esxcli storage core device detached list".

Also "esxcli storage core device list" does not show any storage that shouldn't be there only local disk, iSCSI disk and CD-Rom.

Best regards,

Phil

[a_p_ · Accepted solution]

Just a guess. Maybe this can be resolved (worked around) by - at least temporarily - creating the mountpoint/share on the NAS, and then - once the ESXi host mounted the share - unmount it from the ESXi host.

André

[aerius1]

Yes it is possible this could work as a fix however my current storage are not able to create a NFS share on that exact directory, dumb I know.

So using this method means taking the storage offline and the VM cluster and then get another server up and running with the same IP as the storage with NFS server with the directory ESXi server are trying to access.

Thank you. I will try this method if I find nothing else that works.

Best regards,

Phil

[a_p_]

... my current storage are not able to create a NFS share on that exact directory ...

What kind (vendor/model) of NAS storage do you use? There should be no need to share the exact same directory, the important part is the share name.

André

[aerius1]

It's a freenas setup. As far as I know a NFS share are mounted <IP of share server>:/<directory of share> is it not?

And which means it cannot change the name of the share like an alias and I have to use the same directory as the ESXi server tries to mount else it would not mount.

FreeNAS is only allowing me to mount /mnt/<Datastore> and not outside of that and the share the ESXi server are trying to mount are /data/Folder so I cannot create a share in FreeNAS to that folder.

I have no idea. But if you do know a way please help me.

Best regards,

Philip

[v1ncen7]

I have the same issue with ESX 6.5 and the latest Freenas version.

It is unbelievable that VMware is not able to fix these kind of bugs really. I just decomission the previous NAS. Now I have to set it up again all over just to unmount the datastore from vc??? What a shitty product.

[JochenStucky]

Hi!

we had nearly the same problem. After I found out which Datastore on our Netapp was deleted (I had to make a networktrace to figure this out), it was possible to remove the Datastore with the command "esxcli storage nfs remove -v DATASTORENAME". There was no need to create the old mountpoint. After that, the error message was gone.

Best regards

Jochen

[phreky]

Found this thread on google since I was having the same problem. All of a sudden ESXi was flooding FreeNAS with mount requests, in my case for a share that was very old and hadn't existed for months. I managed to resolve it!

esxcli storage nfs list

showed nothing (after detaching the new share for good measure), as did

esxcli storage core detached list

In my case it was trying to mount /mnt/PersonalPool/Software/ISOs, which had since been migrated (and was mounted and working fine) to /mnt/MainPool/Software/ISOs.

I was unable to convince FreeNAS to create a new share by just mkdir'ing the path back into existance, since it wasn't part of a ZFS volume, it said no-go.

What did work was creating a symbolic link!

mkdir -p /mnt/PersonalPool/Software

cd /mnt/PersonalPool/Software

ln -s /mnt/MainPool/Software/ISOs

After creating the link, The mount went through, and I was able to then detach it from the web interface. I then removed the symbolic link and cleaned up the unneeded folders in /mnt.

[momoraes]

Nice trick! Also, the entry pointing to the "invisible NFS mount" is on /etc/vmware/esx.conf, something like this:

/nas/<NFS VOLUME NAME>/readOnly = "false"                        

/nas/<NFS VOLUME NAME>/host = "<FREENAS_IP_ADDRESS>"               

/nas/<NFS VOLUME NAME>/share = "/mnt/PATH/FROM_FREENAS/"

/nas/<NFS VOLUME NAME>/enabled = "true"                          

I couldn't try to remove it by hand before using the webinterface on my case, but that might work too.

At least you could find the volume name and try to  remove it using esxcli storage nfs remove <NFS VOLUME NAME>

-

MM