Remounted NFS Share. Where are my ISOs?

[TechBrain64]

I changed the IP address of my NFS server hosting ISO content and upon remounting same share and content type (ISO Image), PVE does not display ISO files. The graph shows the consumed space. I have browsed the export and confirmed the files are present.

Is there a method to import the ISOs into PVE?

 

[Chris]

Hi,
are your ISOs located in <NFS share>/template/iso as PVE expects them to be? Do you have the same permissions as previous to the performed changes?

 

[TechBrain64]

Hi Chris,
This is directly from my /etc/pve/storage.cfg file:

nfs: nfs-iso
        export /mnt/Raid10/pve-cluster-iso
        path /mnt/pve/nfs-iso
        server 172.16.5.100
        content iso
        options vers=4.1
        prune-backups keep-all=1

This is exactly the same path(s) used previously. I removed the share, changed the IP addresses (export & pve), then remounted share using these parameters.

Here is the full path where the ISOs are stored on NAS: /mnt/Raid10/pve-cluster-iso/template/iso

 

[Chris]

Hi Chris,
This is directly from my /etc/pve/storage.cfg file:

nfs: nfs-iso export /mnt/Raid10/pve-cluster-iso path /mnt/pve/nfs-iso server 172.16.5.100 content iso options vers=4.1 prune-backups keep-all=1

This is exactly the same path(s) used previously. I removed the share, changed the IP addresses (export & pve), then remounted share using these parameters.

Here is the full path where the ISOs are stored on NAS: /mnt/Raid10/pve-cluster-iso/template/iso

Can you see the isos when navigating to /mnt/pve/nfs-iso/template/iso on your PVE host?

 

[TechBrain64]

Chris,

This is the result when I list the contents of folder:

# ls /mnt/pve/nfs-iso/template/iso
ls: reading directory '/mnt/pve/nfs-iso/template/iso': Input/output error

Is this expected result? Here is screenshot of the graphs:

As you can see the server recognizes the 42.34 GB of ISO data.

 

[bbgeek17]

If you are unable to list directory context from shell, then the PVE wont be able to read it as well.
The NFS protocol queries for "used/free" space differ in level of filesystem involvement from listing directory. It seems you can do the former but not the latter. That indicates an issue with NFS server or local mount.

I am guessing you did not gracefully remove the old NFS mount and its now "stuck" in some form. A reboot will solve it. Or you can try to manually force unmount/remount/etc. Perhaps pointing to a different path in storage config might help as well.

Good luck.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[TechBrain64]

bbgeek17,

I appreciate your feedback on my issue. However, as I've already stated the process by which I made these changes were by the book - as they say - and yet remounting the exact same share, path, permissions, etc. is not yielding what I expected. If this is not possible with PVE just say so. I have grown very weary of this back and forth and no one in any of the other threads I've read has even stated "yes, I have remount a NFS share and was able to access it's contents."

 

[bbgeek17]

I am confident that this can be done with proper planning and preparations. Based on only what you said in this thread I believe the issue happened when NFS IP was removed/changed without properly/fully disengaging PVE.

If you are concerned about infrastructure changes affecting PVE - you can run a nested PVE and test changes there.

If you are dealing with production system, you should have a subscription license and a support case open. If this is not production - reboot will solve it.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[TechBrain64]

If you are suggesting a reboot of the PVE server, no go. I have rebooted the host. Hell, I have even re-installed one of the host and remounted the share. It shows the exact same as the existing host.

Regarding your inquiry about "subscription license", sounds to me you are suggesting the feedback one seeks in the forum is only as valuable as the money you spend.

FYI: I've decided to copy off the ISOs and rebuild the share.

 

[bbgeek17]

I have rebooted the host.

Critical piece of the puzzle. If it was included in problem description, along with I/O error on listing, it would have immediately pointed at your NFS server as a culprit.

good luck

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox