After my vm in proxmox rebooted, the other of 23 virtual server is normal but 1 virtual server that shows only domain id not the real domain as usual. the configuration in /etc/apache2/site-available is still exist but when i try to access web show error 503 service unavailable.
The folder of domain and user are available but when i try to access of feature like file manager is showing error Error opening directory /public_html No such file or directory.
The database menu in the error virtual server also gone and only access in terminal.
If i go to virtual server summary the created virtualserver is Created on 01/01/1970 07:00 AM
How to fix it?, because other virtual server in my vm with this virtualmin is accessible with normal and not errror.
actually i don’t have a backup, because this virtual server account just created few day ago. I’m typically backup virtual server once a week because there is no updated database everyday
I don’t see the solutions actually, the staff asked for backup in /etc/webmin/virtual-server/domains, and yes the file and folder of my virtual server are accessible in /etc/webmin/virtual-server/domains.
If I can access that folder domain, what should i do for next step repairing my virtual server account?
You don’t need to restore the whole domain, just the Virtualmin metadata. It’s actually just one file for each domain that is empty due to a file locking bug, that I’m pretty sure has been fixed in the current Virtualmin version (at least I’m pretty sure Ilia has said it’s fixed, I didn’t fix it or look at the code).
That’s not relevant to this issue, though. The metadata for a domain is in the domain backup.
I’m not talking about other kinds of backups (but I do agree we need to somehow better communicate what “good backups” look like and include, because just having Virtualmin backups is probably insufficient for disaster recovery if you have any customizations in the services Virtualmin manages, since a Virtualmin backup only includes the configuration that is specific to the domain itself…so the VirtualHost section from the Apache configuration, not the whole Apache configuration).
I’m just saying you don’t need to restore the whole domain. I believe you can use the --fix flag in the virtualmin restore-domain command to just restore this metadata, and I think it probably also has an option in the UI.
This absolutely cannot work. The restore-domain command needs a Virtualmin backup file (which will be a tarball) for the --source field, and a --domain that is the domain name being restored. Ilia has linked an example, last one on the page.
If you do not have Virtualmin backups from before the problem occurred, this obviously cannot work. There are ways to fix it without backups, but it’s a lot more trouble (and please start keeping good backups).
The two previous comments from Ilia and I are discussing exactly that option, and that option cannot work without a backup of the data --fix uses to repair the configuration. The restore command restores from backups, if you don’t have a backup the restore command cannot do anything useful.
I can’t imagine how we can make a restore work without a backup. You want us to guess what data would have been in the backup had the user kept backups?
We’ve been working on the issue that leads to config files being corrupted or zeroed out on power outages or other unexpected shutdowns or disruptions of the system, and I believe it’s fixed in the latest Webmin version. (Though it’s still possible to lose data, even if all software is perfect! You need to keep good backups, and if you don’t keep backups, you should expect to have data loss at some point in the future, because things fail. Disks fail, memory fails, data centers burn down or flood, all sorts of things can go wrong that lead to data loss, and if you don’t have backups, we can’t help.)
I was thinking the ability to find a numeric domain if that is the case then try to find any data that is connected to the domain and then try to rebuild it however if the domain id is not added to files, you are correct… your flogging a dead horse
@Joe, I’ll look into creating a virtualmin etckeeper subcommand to manage restores for all or specific domains. This could work by providing extracted configs in a temporary directory or directly replacing actual configs from a backup.
Or, creating a virtualmin etckeeper --domain domain.tld --doctor command to restore all corrupt or nulled files.
even die 
computers are just the beginning (only the internet lasts forever)