What you’re describing is one of the root issues with the current system. It’s the same reason that if your instance goes down, your account and history go with it. I’d love to see an implementation of some sort of account awareness like you said, which could also make it easier to backup history to another instance in the event that your primary goes down.
Oh this would be nice, a standardized way to back up instances so in the case of one going down forever someone else could pick it up and start running.
I know I’m happy to run my instance, I have a great fiber line and a solid infrastructure, but if I get hit by a bus tomorrow I’d want someone else to pick it up and get running
A one-click account transfer to a different instance would be great. However, there can be several “gotchas”, maybe the target instance has lower “permissions”, so that can lead to data loss. Eg: my instance doesn’t allow pictures more than 100 kb, some other instance doesn’t allow creation of communities. So this needs to be carefully throughout.
Seems the easiest solution to that would be to simply have a comparison view. The permissions are setup using some sort of standardized method, yes? Config file, GUI (which just alters config file), whatever. Certainly it wouldn’t be hard to simply grab and organize a list of perms from both instances and toss em up side by side. Could even add notes (i.e. if photo storage on Instance A > Instance B notify user that “migrating to this instance may cause some larger photos to be removed from your account”).
I’m not a professional programmer (sys admin, so mostly just automation) but there are certainly solutions to this.
oh I was thinking a full instance backup. As an instance admin it’d be nice to backup the whole thing in a standardized way so someone else could grab it and spin it up if I collapsed tomorrow, all the community and users
If the instance is setup as a docker container, then it should be easy. The following should be transferred
docker-compose file
zipped up volume directory
At the destination, the docker volume dirs should be unzipped and the new paths should be updated in the docker-compose file. I’m sure someone would have made a script for this by now.
Which is what I’m doing, and I store the backups securely, but then there’s no standardized way of “Hey the admin is gone, can we take over” ability. I could tell my community here are the backups, but then there’s just no standardized way of transferring admin. Idk, I think most things online should have a “Will and testament” type thing. “This passes to XXX”
It’s the main reason for mastodon feeling so off. Subbing to a community is something I can deal with, but having a network where you need to follow individuals and the way of doing it is cumbersome sucks. All of these places would benefit greatly if there was a solution.
I’d be interested in something like a lightweight CDN/replication with OAuth2 for logging into other instances. Each instance ‘replicates’ your original account but isn’t itself the master. One can be promoted to master in the event of an outage effectively migrating your account.
Would make for some difficult security considerations given a rogue instance could attempt to hijack authority.
What you’re describing is one of the root issues with the current system. It’s the same reason that if your instance goes down, your account and history go with it. I’d love to see an implementation of some sort of account awareness like you said, which could also make it easier to backup history to another instance in the event that your primary goes down.
Oh this would be nice, a standardized way to back up instances so in the case of one going down forever someone else could pick it up and start running.
I know I’m happy to run my instance, I have a great fiber line and a solid infrastructure, but if I get hit by a bus tomorrow I’d want someone else to pick it up and get running
A one-click account transfer to a different instance would be great. However, there can be several “gotchas”, maybe the target instance has lower “permissions”, so that can lead to data loss. Eg: my instance doesn’t allow pictures more than 100 kb, some other instance doesn’t allow creation of communities. So this needs to be carefully throughout.
Seems the easiest solution to that would be to simply have a comparison view. The permissions are setup using some sort of standardized method, yes? Config file, GUI (which just alters config file), whatever. Certainly it wouldn’t be hard to simply grab and organize a list of perms from both instances and toss em up side by side. Could even add notes (i.e. if photo storage on Instance A > Instance B notify user that “migrating to this instance may cause some larger photos to be removed from your account”).
I’m not a professional programmer (sys admin, so mostly just automation) but there are certainly solutions to this.
oh I was thinking a full instance backup. As an instance admin it’d be nice to backup the whole thing in a standardized way so someone else could grab it and spin it up if I collapsed tomorrow, all the community and users
If the instance is setup as a docker container, then it should be easy. The following should be transferred
At the destination, the docker volume dirs should be unzipped and the new paths should be updated in the docker-compose file. I’m sure someone would have made a script for this by now.
Which is what I’m doing, and I store the backups securely, but then there’s no standardized way of “Hey the admin is gone, can we take over” ability. I could tell my community here are the backups, but then there’s just no standardized way of transferring admin. Idk, I think most things online should have a “Will and testament” type thing. “This passes to XXX”
It’s the main reason for mastodon feeling so off. Subbing to a community is something I can deal with, but having a network where you need to follow individuals and the way of doing it is cumbersome sucks. All of these places would benefit greatly if there was a solution.
I’d be interested in something like a lightweight CDN/replication with OAuth2 for logging into other instances. Each instance ‘replicates’ your original account but isn’t itself the master. One can be promoted to master in the event of an outage effectively migrating your account.
Would make for some difficult security considerations given a rogue instance could attempt to hijack authority.