[SOLVED] Cannot list PBS backups in PVE GUI

[Bass]

Hello,

My objective is to test the whole backup/restore cycle via the PVE GUI.

I've setup a new PBS 3.1.2, added a datasore and 2 namespaces
In PVE: I created a PBS storage
I could create some CT backups from the PVE GUI. All backups are visible in PBS GUI. Good: backup creation took less time as before PBS

Now i would like to test the backup restore. But unfortunately the backup list is not displayed in the PVE GUI. Here is what i see :

1) It's long loading



2) After a while it shows this communication failure error



Good to know: via the terminal i can list the backups using this command
proxmox-backup-client snapshot list --ns pve



But I am not confortable with the terminal.

I would appreciate if someone could help me to display the backup list in the PVE.

Thanks
Bass

 

[Chris]

Hi,
please verify that the PBS storage is set up correctly on the PVE side. I see you have 2 storages with similar name backup-pbs-pve and backups-pbs-pvf? Are you on the correct storage? Can you see the backups by selecting the storage in the tree view?

Also, what is the output of pvesm status? Further please check the systemd journal for errors, you can get a paginated view of the journal since boot in reverse chronological order by running journlactl -b -r.

 

[Bass]

Hi Chris,

thanks for replying.

please verify that the PBS storage is set up correctly on the PVE side. I see you have 2 storages with similar name backup-pbs-pve and backups-pbs-pvf? Are you on the correct storage? Can you see the backups by selecting the storage in the tree view?

In PBS, I currently have only 1 datastore (called local) with 2 namespaces: namespace pve and namespace pvf
In PVE server, i now have only this 1 PBS storage



I can write backups to this storage. But I cannot see the backups in the tree view. By selecting this storage on the tree view, i receive this error

Also, what is the output of pvesm status? Further please check the systemd journal for errors, you can get a paginated view of the journal since boot in reverse chronological order by running journlactl -b -r.

Output of pvesm status



In systemd journal, I see



I can see some authentication failure.
But I am wondering why the backup list is working in Linux terminal. This Summary is also working well




Thanks + regards
Bass

 

[Chris]

Are you able to fetch the backup contents via the cli by running the following command on the PVE host?
pvesh get /nodes/localhost/storage/backup-pbs-pve/content

How many backup snapshots are there within the datastore? Note that if there are so many, that the request trying to fetch them times out on PVE side, you will have to regroup the snapshots into namespaces.

 

[Bass]

Are you able to fetch the backup contents via the cli by running the following command on the PVE host?
pvesh get /nodes/localhost/storage/backup-pbs-pve/content

Yes, the result is here



The cli seems to work. But the web client not.

How many backup snapshots are there within the datastore? Note that if there are so many, that the request trying to fetch them times out on PVE side, you will have to regroup the snapshots into namespaces.

Just a few number of backups created in the last 2 days - the PBS server is new



Thanks + regards
Bass

 

[Chris]

Please also check the systemd journal on the PBS side for possible errors an clues. Also, please check the response of the XHR requests via the browsers developer tools in the network tab. What status code do the requests fetching the datastores content have? Is there any error messages in the response?

 

[Bass]

Hello Chris,

thanks for all these hints which pointed me in the right direction. Analyzing the XHR requests via the browsers developer tools, I found out that the request had an error. So, in the PVE storage config, I replaced the PBS server subdomain with the IP. The result:

• the problem is fixed, if I enter the PBS server IP
• the problem is NOT fixed, if I enter the PBS server subdomain

I could live with this. But with this info, you maybe know a domain name based solution?

Thanks + regards
Bass

 

[Chris]

I found out that the request had an error

Please share the exact error message, this might help to understand what the issue is. Maybe your configured fingerprint does not match the one of the certificate? Do you use a self-signed certificate? proxmox-backup-manager cert info run on the PBS will provide you with the required information.

Please also verify that the DNS hostname resolution works and points to the correct host. Are there any proxies which might interfere with the traffic in between the Proxmox VE host and the PBS instance?

 

[Bass]

Hi Chris,

Here are my findings:

• the PBS fingerprint was correct.
• XHR Response was empty
  
  

• But the header contained an error
  
  

I did not really understand the error meaning. But I followed your next hint and checked the DNS hostname resolution.
From the PVE cli I pinged the PBS host domain: the IPv4 worked well; IPv6 did not work.

I updated the IPv6 in PBS server (the previous one was wrong)

​

And now everything is working. Apprently the PVE called the PBS using using IPv6. Anyway, the header error disappeared

​

Thanks again for your support.

Best regards
Bass

 

[Chris]

Hi Chris,

Here are my findings:

• the PBS fingerprint was correct.
• XHR Response was empty
  
  View attachment 61639

• But the header contained an error
  
  View attachment 61640

I did not really understand the error meaning. But I followed your next hint and checked the DNS hostname resolution.​

From the PVE cli I pinged the PBS host domain: the IPv4 worked well; IPv6 did not work.​

​

I updated the IPv6 in PBS server (the previous one was wrong)​

​

View attachment 61643​

And now everything is working. Apprently the PVE called the PBS using using IPv6. Anyway, the header error disappeared

View attachment 61641​

Thanks again for your support.

Best regards
Bass

I am glad you could figure it out, please mark the tread as solved for others to find a solution more easily!

I am however a bit surprised and curious why you were able to connect to perform the backups and the cli tooling worked as expected, while the WebUI did not.
Maybe you could you shed some light on the setup: Were these connections from different hosts? I see PVE v8.1 and v7.4 screenshots. Was the IPv6 config done after the initial backups?

 

[Bass]

I am glad you could figure it out, please mark the tread as solved for others to find a solution more easily!

I am however a bit surprised and curious why you were able to connect to perform the backups and the cli tooling worked as expected, while the WebUI did not.
Maybe you could you shed some light on the setup: Were these connections from different hosts? I see PVE v8.1 and v7.4 screenshots. Was the IPv6 config done after the initial backups?

Hi Chris,

yes, I had 2 differents PVE, one v8.1 and one v7.4
Both are connected to PBS in the same way (exaclty the same config; but to differente namespaces)
Both had the same error.

The IPv6 config was done before all backups. This means, with a correct IPv4 config and a wrong IPv6 config:

• creation of backup via web GUI worked well;
• listing of backup via cli worked well;
• but listing of backups (and therefore restoring of backups) via web GUI did not work.

The later point worked only with a correct IPv4 config and a correct IPv6 config.
I have not tested the case without any config yet. But i could do it if needed.

Thanks again and best regards
Bass

 

[jeanphi]

Thanks @Bass

i had the same problem from 2 server on 5 using the same PBS
change IP instead of dns fix the problem

FYI: both are running PVE 8.1.4 on PBS 3.0-2

 

[hexblot]

had the same issue, solved it by adding an entry to /etc/hosts file (on the PVE side) with the IPv4 of the PBS and everything worked like a charm.