Cloudron Forum

Ah great you managed to get this going!

@girish said in Make Cloudron Restore dialog more Human friendly:

Is there any easy fix which does not involved coding?

Yes, have the pattern available in the textbox so it's obvious what is expected.

It's more the "backup date/box ver file" than just a "path".

@andreasdueren yes, usually that means the passphrase is wrong.

Hi @mazarian,

in /home/yellowtent/platformdata/logs you will the file box.log, which should give an indication of what went wrong during your restore.

I did not check what the documentation says about this, but earlier this week I performed a restore in a similar manner to you and watching the dashboard during the restore there was definitely some parallel activity going on during the restore. I would have said that it was probably something like two or three apps at the same time max.

@Fairy this could be a bug in our Import UI, let me test and get back.

As a workaround till we fix:

Download the backup from GCP directly Upload it to somewhere in your server , say at /tmp Then, in import UI, choose Filesystem and give /tmp/app_xxx.tar.gz

I don't think there is any backup option out there which can do what you are hoping can be done. The one exception is if you are manually extracting the data to be restored, and putting them in different locations. I just say this so you don't go feeling like this is some sort of weakness of Cloudron. Backups rely on timestamps and exact conditions. You couldn't restore the second backup because it would look for the time slot from which it was built, in the records, and those records aren't going to be in the currently restored backup. But if it is the media you want, then the fact you do have separate sets of backups does mean you can at least navigate through the backup and find that media, and then as @nebulon said, manually add them back to your Mastodon. But I gotta say, why do you need to have this media? I mean, Mastodon IS social media, and it's meant to be ephemeral. If there is some info you like, then screenshot it, or click the link and bookmark it, print it out to PDF, or follow the poster (to then check them out post-restore to see that specific post) etc. But saving it and keeping it to be viewable only within your Mastodon seems like a use even Mastodon wasn't really meant for.

Anyway, good to hear you are back up and running. Hope the new year remains happier. I'm @shane@social.futurnumerique.com if you want to expand your follow/followers!

Note that when you restore on IP2, Cloudron will adjust the DNS of the apps to point to IP2. It can do this only if you use one of the programmatic DNS providers. If you use manual/wildcard DNS provider, you have to change the IP manually yourself.

@skeats do you need help patching it?

@mdc773 yes. next time around, you can download the backup config that prefills all this - https://docs.cloudron.io/backups/#restore-cloudron

@saikarthik no worries 🙂

So I've tried several times to restore the apps and everything is fine.

@jdaviescoates said in Whoops, installed instead of restored. Do I need to start all over again?:

I'd guess @staff might perhaps add something that will prevent you updating to 7.6 either way if you don't have AVX?

Yep:

@girish said in What's coming in 7.6:

@necrevistonnezr meemo, rocketchat, ackee, wekan & nodebb need mongodb. But the way it's implemented you cannot update beyond 7.5.x . It will just error saying AVX support is needed.

@firmansi You can change the memory limit in a couple of places:

The service (mysql) in Services -> MySQL . Bump the limit there The app's memory limit . App -> Resources.

@3246 very strange, I just tried a restore of Cloudron with SSFS now and it works. Not sure what's the best way to proceed here, but if you are blocked by this, can you write to support@cloudron.io and we can debug fruther ?

Backups are fine now. it created a backup in the night and it succeeded. Only the backup from yesterday is broken according to the log and I still see the old (non existing) backup logs in the dropdown. So just a tiny bug and the backup system itself works fine.

Thanks for confirming this and explaining how it works in the background!

@holm yes, thanks. This is planned but with a slightly different approach. The restore process will keep track of what it's downloaded, so that the download can be incremental. That way, one ideally doesn't need to select anything.

@girish

Thanks, it's the good domain !

@morvin you should instead give the path in the time stamped directories . They will look like hy/2023-01-08-070000-649/box_v7.4.0.tar.gz. There is a version field in the filename. This allows us to check if the backup corresponds to the Cloudron version you have installed.

@girish It worked. It was one of those where I wasn't really sure that would ever work again. So thank you for swift action!

@affinity yes, so in coming releases, we will have a checksum to verify backups. Hopefully, this is caught sooner and not just when trying to restore.