I knew about rclone - I used it as a replacement for the AWS CLI on a project last year. My personal requirement includes syncing with any Linux distro and Windows, and SpiderOak does that well for everything except Silverblue.

It turns out that my Office 365 Personal subscription includes a full terabyte of OneDrive space! I’ve only got about 50 GB on SpiderOak, but since I’m getting the OneDrive space free anyway I’m going to migrate to OneDrive via rclone, which will work on Silverblue. By the way, rclone does not support SpiderOak.

i’m completely new here. since 99 on windows systems and since only 5 weeks the first time linux “experienced”. jumped from one distro to another, because most of the time from day 2 on, only problems occurred. since 5 days working on f29 silverblue and not a single serious problem appeared since then. everything runs more smoothly than under windows or bloated/classic linux - distros.

i’m the living proof that silverblue is beginner-friendly/user-friendly too. it’s not just for nerds, even though it’s fun to become a nerd here. think of me as “the ordinary user out there.” he just wants to do his work on his laptop without having to mutate into a programmer or system administrator.

and thanks to your playful approach @alex285 - i was able to activate auto-updates and more. in the meantime, i’ve also figured out “ostree”, already rolled back and forth, pinned and unpinned, et cetera. it’s child’s play. it’s beauty. it’s full of potential. and i’m ready to build my first own container. thx for this outstanding work.

the only layered package is fedora’s chromium (i can’t stand firefox because i’m tired of the ever - growing user.config, which grows because mozilla makes very questionable decisions & sends data to google & and use too much telemetry ( + one has to deal with 2 policies (google & mozilla) when some “security options” are activated (per default)). and the eye-pain-causing-firefox-gui/font-ugliness under every linux-distro i tried is a complete desaster compared to windows - versions.

since this is the case, one can also use chromium instead and try to “ungoogle” as much as possible. but this is not enough. i would very much appreciate it, if you could integrate a browser that doesn’t phone home (@google & @mozilla itself). fedoras own chromium - fork with vaapi support would be predestined for it. or use the ungoogled-brave browser as standard. that would be much more up-to-date and secure. i’m going into this for so long because i read that firefox should be included by default as flatpak in the next version.

the only linux phenomenon i haven’t understood yet (firefox is standard in every distro - unquestioned). the windows community is a little rougher to mozilla in these days. deserved. based on facts.

but now I’ve gone too far again.

please stay minimalistic and don’t include too much in the upcoming version. what i always had to criticize about linux - distros were the countless applications/packages (i do not need) and uninstalling most of them (on non-ostree-systems) always led to a complete disaster for the whole os.

that wasn’t freedom, that was imprisonment and forcing. something i luckily didn’t have with f29 silverblue. what i need, i get over time. as an appimage, as a flatpak, in a container or if not otherwise possible - as a layer.

what i actually wanted to aks: i have 2 hard disks. one ssd and one mechanical. on the ssd runs (still) windows. because f29 silverblue runs stable and gives better results than windows (also in terms of graphics & performance), this f29 silverblue version should be running on the ssd - drive instead of windows.

is it possible “to save a deployment” and copy it to an usb-stick and “bring it back” to the os after the new installation? or is it possible to “move” silverblues root/swap/home ect. (after windows is deleted) from the second hard disk to the first ssd-hard disk (with any gnome tool or per terminal?).

yes, i wanted to ask that ;). linux-newbie questions. no hardcore developer - questions. sorry for the long text, but there are a lot of additional new impressions.

thx,
noe

There are instructions here (https://spideroak.support/hc/en-us/articles/115001931926-Installing-SpiderOak-ONE-on-Linux-Without-Software-Installation-Privileges for installing SpiderOak One without root privileges. They aren’t correct but I did get them to work by fixing the errors.

this was fast, thx. i take a look at it and try it.

to directly save the current resp. required deployment itself on usb and “put it back” over a new installed f29sb - version isn’t possible (yet … or never)?

You could always layer GNOME Epiphany or install the Eolie Flatpak.

Fedora Silverblue 29 actually slimmed things down quite a bit from 28, and AFAIK that’s a trend that’s largely going to continue.

Assuming you want to save your data on the Silverblue installation, your best bet would likely be to use a classic disk cloning tool such as Clonezilla. Also, you’ll want to make sure the SSD is at least as big as your HDD. Deployments are awesome, but there’s a bit more to it (e.g. the partition layout, boot files) that make it a poor fit to try to use to accomplish this.

hi,

point 1: epiphany or eolie are imho ugly, slower and outdated. i can’t use them. why not brave - browser (it’s open source) as a standard or an “ungoogled” fedora - chromium fork. why stay faithful to mozilla, even though everything speaks against this “company” now. also google is a threat. not even for windows users. when it comes to the web, i see a lot of gaps (even if the os itself is based on security). privacy is a human right. think of stallman & co. .

point 2: yes, i hope one stays with it. everything else would be the opposite of freedom, even if now the whole os doesn’t collapse anymore when you uninstall only pre-installed flatpaks. freedom implies personal choice, without detours.

point3: yes, a backup is imho the better solution. after a look at spideroak, i remembered that i have a tresorit account anyway and can backup certain data here via the browser in a secure encrypted way and without installing anything. but that’s imho not a solution because it’s not a backup in the classical sense. i take a look at clonezilla, thx for the tip.

pojnt4: a pity, i have the impression that deployments would be suitable for such a purpose and i am sure there would be a technical solution to accomplish this for the user. on the other hand, security aspects surely play a role. let’s see for the future. these backup solutions discussed here are not satisfactory. why 3rd party software, if you could also work with in-house tools - based on the architecture of the os itself.

thx,
noe

first it is totally off-topic.

Secondly, Chrome/ium is very unlikely to be your solution to get away from Google. Sure, Mozilla may have made some mistakes in the past, but at least they have been/are transparent about what they do. Unlike Chrome/ium, which now forces you to login with your Google account and whatever they may also do as this projects shows, which you may like. And note it is also about browser (engine) diversity and so Firefox is a god choice, or e.g. Epiphany, which is based on the WebKit (Safari, …) engine. And I’ve never had performance problems or so with Epiphany…
Also note that Brave is also not uncontroversial… In general, I doubt any Chromium-forked browser is any solution. It’s just not independent.

2 Likes

OSTree only really manages your system files in its deployments, even if you backup the entire repo you still have EFI boot files and the fstab layout. It’s not really that simple.

thx refi64 … my system files … isn’t that enough? i need only “my” flatpaks, layered packages, settings of the layered packages, containers (in the future) & home files. possibly not possible at present, but not impossible.

edit: well, i’ve calmed down again. the world isn’t perfect (as far as this mozilla/google - attitude is concerned). there will be some changes. i can wait. on the subject: done. i flattened the silverblue image on the ssd over windows and after the installation i had access to my second hard drive (the previous version) and all its files. flatpaks and chromium had to be installed again. sb on the 2nd hard drive deleted afterwards. rpm-ostree upgrade worked without a workaround this time and the system now boots noticeably faster. auto-updates are set. ready to learn & ready for the future (… without ff & co. - again).

Are you meaning the rpm-ostree? There are ways to refer to your previous meta-data of ostree installs that you can use to get back to a particular state of the OS, external to the Flatpak setup or user data.
First, in a terminal type rpm-ostree status. This will get the commits you have, namely your current and previous. What you are interested in here is the name of your remote, in my case the result looks like below, clipped for brevity


The name of my remote is fedora-workstation:fedora/29/x86_64/silverblue. Using that name in the following command will allow me to “look back” at my previous ostree commits
sudo ostree pull --commit-metadata-only fedora-workstation:fedora/29/x86_64/silverblue --depth=5 five commits. You can make the command option --depth=-1 to go back to the beginning, a lot of info. From there you would deploy a particular commit you were interested in.

wow, thx jakfrost.

yes, i meant rpm-ostree. and yes, i have already “rolled” back and forth & pinned a deployment for security reasons if something goes wrong with my newbie-experiments. also my re-installation a few days ago on the ssd was without problems.

but before this migration i was wondering if it is possible to save a deployment resp. metadata (i.e. the installed flatpaks) on a stick, to upload this data/deployment to the fresh re-installation later. so i wouldn’t have had to re-install the flatpaks (+ settings) and layered package (+ settings) again. a “little” backup (on a stick, not in a “cloud” or whatever), based on the structure of the os itself, without 3rd party tools.

but unfortunately that is technically not possible, i was told here. which is not bad. it was just a thought of mine.

anyway, " … pull --commit-metadata-only …" was completely new to me and is extremely helpful for the future - thank you for that!

If someone manages to Flatpak package Vorta then you will have what you’re hoping for!

1 Like

flatpak create-usb Could be the answer for backing up your flatpaks. I haven’t tried it as of yet but definitely will in the future. I don’t know if you have to do them individually or whether you can bulk create-usb. It looks like you can specify an app/runtime or just the user or the default of the system.

FWIW just to be clear, this does not back up any app data (e.g. stuff in ~/.var), which would have to be backed up separately.

As a side note, IMO that’s where the complications of trying to take advantage of deployments for backups arise, since you still need some more traditional tooling to handle the actual, err, user data. (I’ve had some ideas on modernizing this in an OSTree-/atomic-style way for a long time but haven’t gotten around to it yet.)

yep & thx again jakefrost,

this is exactly the reason why i asked this question at this point in the first place. flatpak -help gave me (on my first day with f29 silverblue) exactly this output: “create-usb -> put applications or runtimes onto removable media”. but i just quickly overread it (i had more important things to do at this point). still something unconsciously stuck. now that you mention that, it’s clear again why i’m asking at all.
so, it is possible. if only sometime in the future. that’s ok. i honestly don’t have the courage to try this option yet.

but even that is not satisfying, because layered packages & containers should also be able to be stored offline on a stick - in this way. firstly, because offline - capacities (against this “cloud madness” known from ms & co.) have to be promoted:

empowering the next billion with ostree, flatpak, ndn, … [edit: but imho not the “cloud” - see r. stallman about this dangerous construct + smartphones/mobile-phones are only a threat in these days; therefore the “monolith” in this video only represents the nsa & co. - these are also facts]

… and secondly because 3rd party applications are imho not a “solution”. there should be the possibility to perform secure offline (mini) backups via the os (architecture) itself (if possible and if this does not violate security aspects).

and i still think that ostree would be suitable for it, even though i can’t come up with a technical solution because i learned something else (unfortunately in this context).

maybe, some ideas (refi) in “modernizing this in an ostree/atomic-style way”, are the future too (just like “flatpak create-usb”). you are all intelligent, passionate people, lightyears ahead of most of the other linux distros. i have no doubt that you will find a solution to this backup - problem (and i perceive it as a problem, also as a security - problem). at least as far as flatpaks alone are concerned, maybe i will play with it (i can always roll back).

thx. and still no problem with sb on my system. it is stable like granit.

uhh: ostree create-usb & flatpak create-usb (e.g. “system”: flatpak create-usb --help ). but a little too cyrptical for me.

I understand that @refi64, it does allow for the backup of the flatpak installation on the system from what I could tell in the help for the flatpak command. That is part of the answer in this case since there is no one stop solution for this problem as of now. Indeed, even if Deja-Dup was working correctly as a flatpak, it still wouldn’t back up the layering and flatpak installations. The layering, ostree, and flatpak states should be able to be captured in metadata I would think since this is the Git of OS’s.

I hope it is helpful, what you are asking for in Silverblue has been asked by others and will continue to be the case until a solution is derived. flatpak create-usb With no options I think will create a usb of your flatpak installations, but I don’t think I have a usb thumb drive large enough for what I have installed. I browsed the app directories with tree, they’re huge. That’s why I think saving the metadata of the ostree state with the layering information, and the metadata of the flatpak’s installed, along with the user data (traditional backup) is the only way to achieve what you are asking. On a side note, but related, I have forgone traditional backups on my personal system for some time, read years. I use the file-roller (Archive Manager) or simply tar.gz the files I want backed up and store them onto an external HDD, the caveat being I have to be organized up front. Git is also another useful tool in this regard, again up front organization is required.

Oh yeah I wasn’t disagreeing, just wanted to make sure it was entirely clear to anyone else stumbling upon this thread in the future.

True, the conversation tends to meander. I didn’t take offense in any case, my skin’s pretty thick. I come from a large family, lot’s of siblings and cousins, and friends with too much time and well developed senses of humor.

1 Like

Is this available as a Flatpak or would layering be required?