Hetzner Storage Box For Proxmox Keeps Disappearing

Software & Operating Systems Linux
,
1

Hey there,
I keep having this issue with my Proxmox setup and just wondered if others have experienced this issue as I’m sure having a Hetzner Proxmox server is probably pretty common.

I decided to take advantage of Hetzners storage box offering as it is a cheap way to get 1 TB of network attached storage for storing backups to.

ksnip_tmp_qbYwct
ksnip_tmp_qbYwct634×643 27.9 KB

I configured my proxmox server through the web GUI (rather than fstab), in order to take advantage of this (over a SAMBA/CIFS interface) as shown below:

ksnip_tmp_zEoVrs

Unfortunately, about once a week, I see in proxmox jobs that the backups failed to run with an error message like so:

image
image1232×602 53.2 KB

When I navigate to the area in the server and perform an ls -alh I see a whole bunch of ? like so:

Untitled
Untitled852×238 34.9 KB

I can easily resolve the issue by simply running a umount operation on that folder and running sudo mount -a which brings everything backup up gracefully. I then manually run the missed backup operation.

I don’t understand why:

  1. This Proxmox server does not gracefully recover from this? I have an older proxmox server on my local network configured with a samba connection to a samba server I set up in Linux, and I can reboot that server without the proxmox having any issues. This was actually the primary reason why I switched away from using NFS, as it would not recover from an NFS connection going away.

  2. Why the storage box appears to be disappearing once a week, or why I have to keep re-mounting. I would think Hetzner would have some sort of better high-availability system than the basic stuff I have set up on my network, and the fact that I can reboot my samba server on my other setup without issue makes me think its more complicated than it just becoming temporarily unavailable.

2

The simplest would be to try with another service - I have used their Storage Box offering in the past and there were all kinds of problems and inconsistencies in the service quality.

2 Likes
3

When Hetzner sees Plex activity, it is being told by Plex to disable access for legal reasons. Maybe something like this is going on?

4

Okay, that’s good to know. We will look for a different service and see if it is just specific to that service.

5

No, plex isn’t on any of these servers.

6

If connection drop due to some maintenance (which often happens) the connection will hang and the kernel won’t try to recover.

People on their forum seem to recommend autofs.

If it’s a backup job I think it’s better to mount the volume right before the job and then close it. Or use something like scp.

1 Like
7

I’m going to try first with using the systemd automount option using the following first with a manually configured mount in fstab, and see if that one remains fine. If so, then I will just switch to using that. I would rather not have to rely on setting up scripts to manually mount a volume just before backups and then unmount later.

//xyz.your-storagebox.de/xyx  /path/to/mount    cifs    credentials=/path/to/creds,uid=0,gid=0,_netdev,x-systemd.automount,x-systemd.idle-timeout=60s     0     0

I’m just confused as to why I’m having this SAMBA issue in the first place. I found that I had the hanging connection issue you described with NFS, but not with samba on another proxmox setup I had (but which was using a much older version of proxmox).