[SOLVED] [Unsupported] NFS doesn't work on PBS

[enoch85]

I'm giving up soon. Now the NFS share works perfectly fine but PVE can't find the datastore:



I don't know if it's a pattern, but it seem like the connection to the datastore in PVE is dying when rrd is run. I might be wrong though.

Dec 04 17:28:38 pbs proxmox-backup-proxy[1176]: rrd journal successfully committed (25 files in 0.027 seconds) Dec 04 17:49:13 pbs proxmox-backup-proxy[1176]: backup failed: connection error: connection reset Dec 04 16:52:06 pbs proxmox-backup-proxy[1177]: starting rrd data sync Dec 04 16:52:06 pbs proxmox-backup-proxy[1177]: rrd journal successfully committed (25 files in 0.041 seconds) Dec 04 16:53:34 pbs proxmox-backup-proxy[1177]: error during snapshot file listing: 'unable to load blob '"/mnt/qnap/vm/110/2022-12-04T15:04:20Z/index.json.blob"' - No such file or directory Dec 04 18:06:20 pbs proxmox-backup-proxy[1166]: applied rrd journal (242 entries in 0.037 seconds) Dec 04 18:06:20 pbs proxmox-backup-proxy[1166]: write rrd data back to disk Dec 04 18:06:20 pbs proxmox-backup-proxy[1166]: starting rrd data sync Dec 04 18:06:20 pbs proxmox-backup-proxy[1166]: rrd journal successfully committed (20 files in 0.022 seconds) Dec 04 18:07:43 pbs proxmox-backup-proxy[1166]: error during snapshot file listing: 'unable to load blob '"/mnt/qnap/vm/130/2022-12-04T00:12:19Z/index.json.blob"' - No such file or directory (os error 2)'

 

[Neobin]

As far as I followed this, all your tests are with NFS. Did you already test with SMB too? At least I am wondering how it looks like with SMB in comparison in your setup...

 

[enoch85]

@Neobin Haven't tried that, but been thinking of trying it.

Right now I re-added the storage in PVE to use the DAC IP instead as the management IP seems to die from time to time. So far so good, and memory (RAM) is fully allocated.



It doesn't seem to utilize the 10 GBe connection though, but it can be due to RAID5 on the QNAP, super slow.

 

[enoch85]

MISSION SUCCESS! (I think)
OK, so the long journey is over. What the issue was all along was DHCP on OPNsense, at least that's the most logical explanation. Beause when I deined PBS DHCP through MAC filtering, it (PBS datastore in PVE) stopped timing out - and besides that, even if it did timeout before I made the DHCP deny, the transfer continued as it's not at all linked with any switch or network other than the /29 net I set for the DAC between QNAP, PVE and PBS.

So the final setup became:

1. Jumbo frames (9000) on everything - QNAP, vmbr3, eno49, the 2 OVS ports (1 per PBS/PVE), and also inside the PBS configuration since it got a separate NIC.
2. Reverted back to this in the /etc/fstab mount: 10.255.255.10:/PBS /mnt/qnap nfs defaults,proto=tcp,nfsvers=3 which gives me this when typing mount on the PBS: 10.255.255.10:/PBS on /mnt/qnap type nfs (rw,relatime,vers=3,rsize=262144,wsize=262144,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=10.255.255.10,mountvers=3,mountport=30000,mountproto=tcp,local_lock=none,addr=10.255.255.10)
3. On the QNAP it stays the same as my last edit.
4. Permissions of course, handled on page 1 in this thread.

I'm satisfied, even if it's slow, since it's just the first backup now that will take a full day, rest will be quick.

Sorry to all for the nagging, it's easy to blame when all you see is a dark tunnel in front of you with no solution. I hope this thread will help someone struggling like I did the past week.

Pic or it didn't happen:

 

[enoch85]

Small update, it's still going since Sunday, so all good on that front.

Still having issues with SSH and Web-UI access (to PBS) dying on me after some time, haven't had time to track it down yet. Any suggestions?

 

[supersic]

Hi, I'm trying to follow along with this thread but can't even get connected to my Synology device.

My current setup is that I'm running PBS on my Synology in VMM with 100GB assigned for the virtual machine, and then I'm trying to backup to NFS on the same device, but outside the virtual machine. My PVE is on a separate device altogether. I'm able to backup to PBS, but only within the PBS VM and can't mount the NFS share from PBS (but the NFS share works from PVE).

When I type 192.168.86.39:/volume1/ProxmoxBackupServer /mnt/synonfs defaults,proto=tcp,nfsvers=4:
I get the response:

-bash: 192.168.86.39:/volume1/ProxmoxBackupServer: No such file or directory

Not sure where exactly I'm going wrong.

 

[Dunuin]

You could try it with nfsvers=3. Also keep in mind that PBSs "backup" (I think that was UID 34, run cat /etc/passwd in PBS to check it) user needs read/write access to that NFS share. And the root user (UID 0) might need read/write access too.

 

[Neobin]

When I type 192.168.86.39:/volume1/ProxmoxBackupServer /mnt/synonfs defaults,proto=tcp,nfsvers=4:
I get the response:
Not sure where exactly I'm going wrong.

You either have to let it mount automatically via the /etc/fstab file: [1] or manually with the mount command: [2].

Edit: Pasting only the parameters in the shell without a command can not work. (At least this is what I understand from your output, that you did try.)

[1] https://wiki.debian.org/fstab
[2] https://manpages.debian.org/bullseye/mount/mount.8.en.html

 

[AndreCast]

Hi,
If I have a PBS with a 1GB network to which I connect a 10GB NFS storage, will the backup of the VMs (10GB cluster) use the PBS network card (therefore 1GB) or will it reach 10GB?

Thank you!

 

[Dunuin]

Backups from PVE to PBS will he done using the same interface the PBS webUI/API is using. So you will be limited to 1Gbit.

 

[fabian]

it depends on what you use as host on the PVE side and how that is resolved/routed. so if your PVE host(s) also have access to the same 10G network, you could set it up so that it goes over the link for that instead of using some default route over 1G.

 

[NinthWave]

OK, so I FINALLY got it working again by adding a new user and group to QNAP named backup, with UID and GUID 34.

[~] # addgroup -g 34 backup [~] # adduser -u 34 -G backup backup [~] # Changing password for backup [~] # New password: [~] # Retype password: [~] # Password for backup changed by admin [~] # cat /etc/group administrators:x:0:admin everyone:x:100:admin,enoch guest:x:65534:guest backup:x:34:backup [~] # cat /etc/passwd ..... backup:x:34:34:Linux User,,,:/share/homes/backup:/bin/sh

The problem with this solution is that QNAP allows one to changer userID and groupe.
One cannot do that on TrueNAS core...
Do you think of a workaround ?

 

[Dunuin]

The problem with this solution is that QNAP allows one to changer userID and groupe.
One cannot do that on TrueNAS core...
Do you think of a workaround ?

Its possible to change the owner of the NFS share to UID 34. I'm also mounting the NFS share from TrueNAS Core to my PBS.

 

[NinthWave]

Its possible to change the owner of the NFS share to UID 34. I'm also mounting the NFS share from TrueNAS Core to my PBS.

How? Did you change that using ACL or from within Sharing, NFS, Edit ?

 

[NinthWave]

Its possible to change the owner of the NFS share to UID 34. I'm also mounting the NFS share from TrueNAS Core to my PBS.

Am I missing something because:


even if

root@pbs:~# cd /mnt/pbs_core
root@pbs:/mnt/pbs_core# ls -la
total 12014
drwxrwxrwx     3 nobody nogroup     4 Jan 30 21:40 .
drwxr-xr-x     3 root   root     4096 Jan 30 21:37 ..
drwxr-x--- 65538 backup backup  65538 Jan 30 21:42 .chunks
-rw-r--r--     1 backup backup      0 Jan 30 21:40 .lock

 

[Dunuin]

Can't look at the configs right now as the server is powered down. I think I had to work with the "mapall" or "maproot" option of the NFS share.