flux

Kopia has served me great. I back up to my local Ceph S3 storage and then keep a second clone of that on a raid.

Kopiahas good performance and miltiple hosts can back up tp it concurrently while preserving deduplication -- unlike borgbackup.

I believe you're completely right here, except that snapd can be configured to point to another store, though it's not very well documented.. I did find the piece of information once :).

But the thing is that the client still only supports one app backing site at a time. So if you pick another one, you lose visibility to the other store. I doubt even updates work as they should.

So it's really about building technology that is geared towards centralized control, whereas basically anyone can host flatpak packages and give ref links to them.

As I understand it, these kind of applications depend on being able to perform activities in the background, which is highly limited in iOS for battery efficiency reasons--and maybe for privacy.

Many years ago I was working on a project that shared connectivity details over wifi/bt, and iOS was troublesome also due to the application not being aware of the local bluetooth address.

Possibly similar issues impact other mesh networking applications on the platform.

I guess they can be pretty safe from radio interference there, at least :).

I doubt the connectivity issues need to exist, though, probably works just fine in some configurations. What I'm wondering though if they had a spare, and maybe a second spare, and space batteries, on the boat. Or possibly manual override (doesn't sound like it).

I think the device itself is fine, though it might be indicative of too aggressive cost cutting measures.

I also requested perhaps a bit more than two weeks ago and got it a couple days ago.

Used https://github.com/xavdid/reddit-user-to-sqlite/ to put it into a more structured form. I guess I should give Datasetts a try to easily browse it, the project's README links to it.

You can use the web ui remotely.

Personally I use it from command line, though, and my only complaint is that it's too easy to start a backup you didn't intend to.. Buut if you're careful about usong the kopia snapshot command then it's fine.

Depends on how much you change per time unit.

I take full system backups every three hours, but the backups are thinned so that there are previous 24 hourly ones, previous n daily ones, previous m monthly ones, etc. Similar approach can be used with snapshots.

I don't currently use snapshots—I don't run btrfs anymore—but when I did, I did a snapshot every hour and kept them for 24 hours. But then I backed up the latest snapshot, which gives consistent backups, versus regular backups where files can change while you're doing them. I'm nowadays using bcachefs, but I don't quite trust its snapshots yet so I haven't started using them ;).

This would be bigger news had they broken WA E2EE. Indeed, the officials might prefer not to disclose the capability if they had it and this wouldn't have happened. (Except, maybe, via parallel construction.)

You will reconsider calling strategy a backup should the filesystem get corrupted for whatever reason.

I've tested my full system backup restore once with btrfs. Worked out fine.

Too bad, would’ve considered it as a viable option to mdadm + BTRFS.

Currently I'm using bcachefs with LVM (which can do raid, but I currently only have one NVME SSD), though it indeed does have RAID1/0/10 support. But overall I expect it not to not make the same silly default choices as btrfs, such as not being able to start the system if a RAID1 component of your root filesystem is missing. And, supposedly, when the RAID5/6 becomes stable, it won't have the write hole problem.

It said the code base was build on something stable, but it didn’t say what, do you happen to know what FS this project is a fork of?

It's based on bcache :) by the same author, but of course bcache is not really a file system but rather some kind of object storage layer for the purpose of caching slower block devices and absorbing write load.

Bcachefs might be coming soon to the mainline kernel, so that's going to make it a lot easier to try out. Personally however I have lost one bcachefs (that FS was readable, though, and I have good backups), but I have also lost a btrfs before and seen reiserfs bugs, so I don't too heavily count it against it; overall I enjoy its stability when using basic functionality. I haven't dared trying snapshots with it yet..

They also have a newer model Astro Slide, but it's significantly more expensive and has availability problems. I've had one since Christmas or so, though, and I'm happy with it. CC is fine as well, and indeed the sale seems quite good.

(I used to have their first phone as well, but truly AS would have been a much better first device to bring to the market, there really aren't any fundamental flaws in it in my view.)

Another way to check is to

strace cp testfile testfile2

and the sequence in which the message is printed and operations performed can be studied.

It's perhaps a lot to read, but linux tracing tools are worth learning!

I do think the idea behind snap isn't all about pushing the Linux platform as such forward, but to specifically gain a market advantage to Ubuntu.

Why else is finding documentation for changing the default store so difficult? And I don't think you can even have multiple "repositories" there--quite unlike all other Linux packaging systems out there. (Corrections welcome!)

I guess currently it doesn't seem that great, given they've been unable to fulfill their Indiegogo campaign due to manufacturing woes. I hadn't realized how lucky I was to receive this unit :).

I do hope they'll survive, though, because I want to keep seeing qwerty phones—and have some support for this one..