All this wen fine, the problem is about restoring the domains.
I started with one domain from a full backup.
The domain validation says that can’t restore apache related configs, but runs to the end.
For some reason, after restore, the new domain doesn’t appear in the virtual server list.
If I try to restore again, it complains that the virtual server already exists.
What is wrong here? How do I force it do show the newly inported domain from backup?
It wasn’t enough, it then complained about existing user.
So, to avoid the cleanup procedure, I’ve re-installed the server again and created a snapshot of it with virtualmin installed only.
This way I can easily revert to this point.
So after this, I’ve tried again restoring one of my domains from the LAMP server backup, and once again, it didn’t add it to the virtual server list.
How can I debug this ? I need a way to import my servers if possible.
Help appreciated, thanks.
One additional note.
I’ve installed this server with XFS file system, and it complains that can’t enable quotas.
I couldn’t find it where to remove, so just re-installed.
I’ve the snapshot now, it’s easy to revert.
Can you help regarding the domain not showing?
One thing, I think the process doesn’t finish, because I don’t get the return button in the end of the import process.
Could this be related to php timeouts ?
I don’t think there is any support for restoring an Apache domain into nginx. I don’t see how there could be, as the config files are wildly different and nginx lacks a ton of features Apache has.
I’m not sure, honestly, it’s not something I’ve ever done. And, yes, not having a web feature is gonna choke Virtualmin on restores. We never anticipated not having or wanting the web feature (though in this case it’d make sense to restore everything except web and then enable web after with nginx and modify the settings manually to suit the site in question). There is probably some way to trick it into doing most of the work, but I don’t have a strong sense of how.
I talked to Jamie about it, mostly just to ask him to make the error much more clear that restoring from Apache to nginx is not possible, and he responded that maybe he’d be able to make it possible. So, maybe it’ll show up in limited form in a future version…but, no promises. It’ll never be a 1-to-1 mapping because they’re very different servers and nginx just can’t do a bunch of stuff Apache can, but for a lot of folks, it’ll be sufficient, and at least would get you through the “can’t get there from here” problem you currently face.
One avenue you might try is installing/enabling Apache (disabling nginx), doing all of your restores, and then converting the server to nginx and manually fiddling with the web server bits. I would never suggest this in normal circumstances, since it’s a bunch of manual steps to switch between servers, but it’s probably less work than restoring all the pieces manually.
Jamie said it is expected to work (with obvious limitations based on them being very different web servers with different capabilities), so the problem OP reported is bug-like, but we’d need to get more info to fix it.
I’ve finally found a way to proceed.
So since the exclude-feature didn’t work, I’ve tried dissociating the apache and ssl features from the domain.
This of course causes a downtime, the webpage stops.
At this point I do the backup.
when complete, I associate the features again in prod server.
Then for restoring, I do a normal all-features restore, and this time it works!!!
I’ve the domains imported, but just with no web features activated.
I’ll now proceed with tests, I’ll reactivate web features in test LEMP server and see if I can get the webpages working.