been running rsync/rsnapshot for like ~ 20 years on server. dropbox and syncthing for desktop/workstation. plus git for projects. never needed much of a desktop backup solution. but i recently started setting up linux desktops for friends & family TM. pika looks like the right tool there. thanks for creating great content!
Right with you, after years of half-assed backup strategies using DAT and CD/DVD-RW, I setup a dedicated rsnapshot box on an SFF system. Dedicated because I realized that I was the primary danger to anything mounted on my primary machines. Quarterly backups to Nov 5, 2004! Admittedly, rsnapshot is not so user-friendly as this tool, but in exchange I can use Linux knowledge to accomplish useful tricks, like moving to bigger drives, mirroring to an offsite archive, and fixing things when something goes awry.
I'm curious, do you encrypt your rsync/rsnapshot backups? I usually make a tar.gz file, encrypt it with gpg and then scp to a server. It has the advantage of being seamless (gpg asymmetric encryption does not require a input), but that does not make incremental backups
@@nogesax3535 In my case, I have a local rsnapshot server, 2 external mirrors (rotated off-site once a month), and a Google Compute Engine mirror. All backups are on encrypted filesystems (via LUKS). Encrypting at the file level would be somewhat opposite of how rsnapshot works - not saying it's not possible, just that I worry that it would be hard to manage successfully and reliably while also not leaking data.
in my case no encryption. external hard drive. same disk for almost the full time. replaced once to upgrade size. old WD "Book". so incremental works ofc. backups usually take only seconds.
This is one of the very best linux software tutorials I have ever seen. Very detailed but NO fluff and unnecessary "background music". Well done. earned a like and a sub
Anyone who has hearing loss has a lot of difficulty separating the music and the narrative. That is my rationale for disliking "background music". The second reason is that some folks do not know how to level the audio narrative so that it is consistent across the records so you constantly have to tweak the volume control up or down to be able to understand the narrative. Thanks for asking! @@wesbryie
@@tennisfreak312 Same problem that I had when it came to backup-ing, that I switched back to Windows after almost breaking my computer while trying to restore my system... 😅
Not only the topic is interesting. I'd like to acknowledge separately the level of video production with those blur and highlight effects. Does feel classy comparing to our usual YT Linux stuff videos (not that those are necessarily bad, but...). Geez, an instant subscription.
in kde systems there is KUP, which makes incremental backups and is integrated with the system itself. The behaviour of KUP is similar to file history in windows, it takes backup on the specified interval and auto detect if the specified drive is mounted or not, if not then it will not do anything and as soon as you connect your drive it automatically take a backup. the ease of use of KUP is identical to file history in windows. You can access kup in system settings in kde and setup multiple schedules if you like
quick tip to pika devs: if you're targeting user friendliness for the interface, maybe don't say 'repository' and just say backup location. not everyone is a developer and knows what 'repository' means. this could cause some mild confusion when they click backup but are then confronted with a question about setting up a repository and going, "huh? wait what's a repository? I just want to create a 'BACKUP' "
100%. I'm a completely new Linux user. Been using it for one week at most. Prior to this, my experience with GitHub amounted to knowing its name. I had also never heard the term 'repository' before, and I'm still somewhat confused on what it means.
vorta is a python app. pika backup is a rust app, that means it takes a long time to compile using AUR in Arch Linux :( in my test VM i enabled chaotic AUR for fast install aur precompiled stuff
I have nothing setup at the moment and this is exactly what I was looking for. A life saver I must say. I also love the open nature of the tool, that it has a simple side and also a full on complex side.
I think it is because, the first Video is only a Month ago. It needs a bit of time to gain subscribers. But I think for the short time, that this Channel is active, it gains a much of subscribers. But this Channel needs more subscribers. I already love this Channel
*It's a real shame you can't find this great video if you google for Pika Backup* - or only if you're lucky. A title like this would be much better: *Pika Backup: We finally have a simple solution for backing up files on Linux*
Hey just to let you know that Deja Dup Backup also existed for a long time and similar to this but this tool looks very awesome and probably way better/polished and advanced.
I'm a simple user, so this might be too simple, but I only use `rsync` and a bare git repository to backup my entire machine. A post-installation script will restore all of that along with installing packages in 30 minutes, I can reinstall my machine right now without the fear of losing anything.
For the longest time I was just stuck with using a mess of an rsync shell script to have some resemblence of version control and such. Been tempted to get into borg but the barrier of entry felt very daunting. Pika on the other hand looks amazing, need to try it out ❤
Let's be honest, the user experience for Pika (like a lot of gnome) is very nice indeed - Linux has always been powerful, but now it's starting to get the whole design/ux piece right. As a recent convert, I've been messing about with snapper and btrfs assistant - I'll have a look at Pika :) Awesome video, thank you!
Honestly, I'm really vibing with what Gnome is doing at the moment. The level of consistency and ease of use I think is going to push the platform higher when it comes to mass market. It's almost at the stage where I could start to recommend it to non technical people.
I am using Vorta and it served me well - but pika looks a bit more polished for "simple setup and every day usage" for people who don't want to understand the inner workings of borg. I am happy that you covered it - I will just send anyone I find without a borg backup to watch this video in the future
At present, simple does it. Straight forward bash scripts to rsync to a couple of ZFS repositories with snapshots before and after. It may not be the most elegent, but easy to understand, and has got me out of trouble more than a few times.
I would not install or use z. FS on an envy M. ESS d until they fix the piss pour. Random reed and writes performance out of the box by default From what I understand there isn't simple file. Or command that 1 must edit. I think involving something about highsync or usync =13 ...to fix it? I'll just stick to the tried and true ext4
@@motoryzen People don't use ZFS for top performance, they use it for data integrity and/or its management features. For what it's worth, ZFS works fine on my NVMe machines (not "piss poor" as you might say), but of course it'll always be slower than ext4 or XFS simply due to the extra checks that ZFS does.
@@austinleong3319 yes I know what they use it for but the bottom line is the random reading right performance is so shity compared to the normal standards that we've been used to with nvme technology that it basically defeats the point One and even better analogy to see what I mean go back to dial-up internet
Auf den Punkt. Kein Stottern. Keine "Ääähs". Keine "Eeers". Einwandfrei gemachter Beitrag, mit dem ich meine Zeit nicht verschwendet habe. Von der Art Vorstellung will ich mehr sehen. You are subcribed! :)
Used mountable sector-wise backups of the whole partitions: zerofree+dd+mksquashfs with pseudofiles compressing on the fly. Used zpaq for multiversioning of one directory. Not especially user-friendly but achieves maximum compression+multiversioning+deduplication.
timeshift does just fine as a backup solution and works like time machine, only with more user configurable options. It's meant just for system backups, but you can include user files as well (or just your dotfiles). Backintime is also pretty good. Fully featured and nice GUI. I don't understand how this fails the simplicity test. Configuring timeshift takes about 10 seconds with the GUI client. Not saying Pika is in any way bad, but Linux has not been short of backup solutions.
Not a lot of mentions (if any) for restic, it’s not as simple as Pika, or rsync but I’ve found it fantastic for backing to object storage, it does it using packed blocks and has a repository file saving non-negligible amounts of money with transaction costs especially when backing up source code. I took a look at Borg a while back and couldn’t get my head around it for my use case (full disclosure I may have not tried hard enough). But yeah restic has been awesome, backing up multiple hosts efficiently and recovery has been straightforward. Worth a look I think.
Oh just to add, it has a sparse checksum validation feature as well to do integrity validation on random subsets… big win when you have a large backup set.
Using restic for my homelab servers. One local hard drive repo receives the files from all the servers. The repo is then synced off-site with rclone. All that process is automated via a few basic scripts. Pika looks very nice for a desktop use case! I wish there were the same UI for restic so I could add a desktop backup to the same repo I already use (though I have limited need for desktop backup, using a dotfile git and storing files on the servers rather than locally)
@@bernardcrnkovic3769not them, but not sure, there is also restic-rest server. I suppose in theory it's supposed to be faster. It would mean going through less protocols, too. That is where some inefficiencies can happen with chunking between protocols. I'd be curious in what people find though
Thanks, I was looking for backup solutions for a new Linux installation. One thing though, you don't need to use the time machine interface to browse files, you can just open the backup drive and access one of the time stamped backup sets.
Nice, good improvement over other options. That said, it's way more fun to setup another system as a server, install a RAID drive array, then mount it as an SSHFS mount, then write an rsync script to backup everything I want, then run the script via Cron job. I don't do things the easy way :-)
i use a script that uses borg to backup my markdown notes to my USB stick sitting on my fritz router mounted as cifs that way if my hard drive dies then at least my notes are in a more current backup state than the real backups i don't do that often to external USB drives
Pika Backup on Linux-mint and presumably Ubuntu, 760Mb download, 2.6Gb on disk (Flatpak). Now, I don't know a lot about computers but that sounds mindblowingly bloated to me for something that I could fashion with rsync and a bash script in just a few dozen bytes.
Was thinking of Borg right from the start, having used it for the past 6-7 years, and lo it turns out to be a borg GUI! Pika looks like a pretty good tool to get to an initial borg config even for the cli-inclined.
Clean video production! I like Pika Backup, but since full system snapshots aren't really a thing on Linux (other than Timeshift and btrfs assistant which required arduous tinkering to set up), I just back up my entire Home folder in Dropbox in an encrypted .zip file, along with a post-install script that installs everything on a fresh install.
I use my own shell script run as a cron job. It backs up to a selection of USB thumb drives. For each type of backup (home, /usr/local, mail, /etc) it: * deletes all but the latest N (usually 2) copies by date from all backup drives, * creates the backup in a temporary space, * copies it to multiple thumb drives (usually 3).
I’m glad I bumped into this! Now all I wish to try is, hope it can backup my whole distro and cache files incase if I wish to to move to another distro like popOs to fedora! Gonna check this out. Have gladly subscribed ! Thanks for the video 🎉
I've been using syncthing while I was developing my own solution for backups. After seeing this and giving Pika a shot, I'm happy to say that I will no longer be hearing my laptop fans spin up and my computer lag many times a day as syncthing does its syncing.
I'm using Synology Drive. I don't backup whole system, just documents and project files - source code, images etc. It works really well for me. For backing up applications, I created sh file which will install all apps (flatpak and deb packages) on new machine. I don't backup whole system, because I build large projects - linux kernel, PrusaSlicer and other SW, so build deps would take a lot of space for backup.
This is a wonderful, and timely video, (for me). I've been looking for something like this since I first started using Linux years ago. I'll be looking forward to setting this up on my system and experimenting with it. It really looks promising. Thanks for making this video. I'll be looking forward to your future videos on the subject. /Bob (A Now-New Subscriber)
@@b43xoitAnd that "error" would never have been solved, while rsync is as old as Linux is and is the basis of about any back up system out there? Hard to believe.
Yeah, I also went with Timeshift for the System and PIka for the personal files. Timeshift saved my ass every-time I tried something crazy enough to destroy my whole system.
Pika Backup is a frontend for Borg Backup. As used with the CLI you also can use Borg directly. Not hard at all, has a lot common with git. Instead of Pika Backup i now used "Vorta Backup" instead. Vorta also is a frontend to Borg Backup. I picked Vorta because it already ships as a Debian default package, so no flatpack needed. But Borg itself is really easy to use, so in general great option! Used Deja Dup before but it somehow is slow in backup and restore. Also one time my backup was broken and it just told me that everything is lost and I need to rebackup from start. Usually not what you expect from a Backup program. A Backup program should be usuable even if the backup itself is broken. At least Borg seems to provides a check, repair and also provides option to use it even with broken files. Its also extremly fast. Good for Desktop backup. For more enterprise stuff "Bacula" is a very good option. Used it before and replaced it by horrible rsnapshot.
Honestly one of the reasons I love openSUSE is how easy snapshots are with it. Yes I know, snapshots are not backups and I do have my important stuff secured elsewhere but for the most prevalent problems: Broken updates or breaking things with configuration, it's enough.
my system backups were living on a prayer. if I broke my system, I would have to manually plug in a live USB with linux, mount my laptops drive and partitions, and search for the files manually (hoping they are still their). Now that I know of this tool, I have already started downloading it while watching the video :)
Thank you so much for a very good review. Very helpful, indeed! I'd be grateful if you can top this off with a video on Borg backup some time in the future!!! Keep up the good work, your channel is amazing!
TimeMachine is one of the most important reasons I still use OSX/MacOS. It has saved me twice, once restoring from to a completely replaced drive. Not just my files but my full OS including all settings for printers, wifi and everything else. It is simply brilliant, I really don’t understand why Windows and Linux don’t have something similar. It is from 2007, so it completely baffles me that no other OS has been able to create something similar in 17 years..
as a regular non-tech-savvy user, when it comes to backing up personal files, it's rather simple and easy. I know which folders to backup. I was wondering more about what should be backed up for reverting a system back to it's previous state after a system update that went wrong, or something like that. But I don't even know whether that's possible to do on a mutable system. I guess I should search for some tutorials about it.
If you just backup your entire root directory you should just be able to copy it back and be fine, at least if it's all on one physical device. If you have a weird partition structure with other drives mounted on your fs then you might need to recreate that manually before you restore. The only issue is your /boot volume, whether it has the right kernel image for the hardware you're running on (if you need to switch to another machine) and the boot loader / uefi partition. But if you run a linux install disc with minimum options and then just copy your files back that would probably be fine. Just exclude /boot from the backup. Now you can make it harder if you want to actually save space and not back up things like package files, but that can get complicated. If you're using an arch based flavor I can tell you that /var/lib/pacman stores metadata about all its packages, including file hashes. But you need to write/find a C program that calls libarchive (with gzip and mtree extensions) to read it. Or you could just exclude the output of "pacman -Ql" or whatever command lists all files on your system that comes from packages (apt list?). But then you'll lose anything that you modified, like /etc/mirrorlist
I'd just like to simply copy data - no repository BS, no "restore later" nonsense, no compression. Just copying over and check for duplicates (and overwrite them). I want to backup my data for use later, not to bind them forever onto Linux/UNIX. This is a nice tool for a crappy backup logic behind (borg).
As an openSUSE Leap, KDE. user for years now I will see if I can apply what this Excellent and helpful video has shown me, to my system, using a spare Crucial MX500 sata ssd. Even having some knowledge of the CLI, the prospect of Linux backups has, until now, seemed off putting . Great work. Thank you and subscribing.
If you cannot select a certain filesystem for formatting, that means you do not have that filesystem installed (such as ntfsprogs or exfatprogs). for example, fat and ntfs are often installed by default, but f2fs is less often installed by default. There are options under Other that may be available, it just doesn't show them on the first backup page.
I'm looking to implement a strategy where I sync all my persistent data (documents, music, etc.) to my NAS via Syncthing and then just do the backups server-side. Since I use NixOS, any system configuration can easily be reproduced with a few commands. I'd only really miss out on backing up "temporary" data like local container volumes, browser cookies, downloads and maybe some saves from Steam games.
Agreed! I've been playing more and more on nix on my main workstation and I think Pika/Borg goes nicely together. You can also manage Borgmatic from your nix configuration as well.
Since I'm on butter I use Timeshift to snapshot my filesystem hourly. Since I don't really have anything important on it, I don't bother backing it up externally. Everything of importance is on my NAS and modern SSDs are reliable enough for me to not bother as much
Thank you for this. I've been looking for a BU app that was easy (so that I'll use it) and also writes accessible files to the archive. Having been on Retrospect for many years, I was always concerned that the archives were not even mountable on any OS, let alone retrievable. It seems the only thing missing from Pika is the last item on my list: I wish it was multi platform, or at least on Linux and Windows. Oh well.
I use a combination of Syncthing and Duplicati. Syncthing is very useful for network drives or remote PCs. Then I use Duplicati which is installed on a docker container on my Pi5 to do encrypted backups to my NAS that is running in raid 1 and another in raid 6. 😌
What micro computer are you using? Would you recommend? I’m in the market for something besides a laptop. I don’t play games, I’ll run a nice linux wm, and do development. Thanks.
I have a Raid of 2 8TB drives with all movies, music & photos backed-up, and all family members can get to them. I have another Raid of 2 8TB drives, and use a shell script I wrote to backup weekly.
Thanks for the great video! I have one question, if you don’t mind. Would this be a good solution for a frequently detached external drive? I’m planning to use Picka to back up my personal laptop every weekend, for example, only when I connect the laptop to the external drive.
Looks good, but I've been searching for incremental backups. We used to do a Full one week, three Incremental weeks, and keep the last two months. This system was trade off between space and finding files. You only had to look through a few Incrementals to make sure you had the latest file.
I had some issues with Pika that made me switch away. It would only work in Gnome, so I would be unable to access my data id using a different desktop. I would also get issues after mounting a lot of backups in one session, so I would need to close the app and reopen it. I switched to Vorta instead, which didn't have those issues. It even shares the same backend, so my Pika backups still worked with Vorta.
I don't really need a backup solution for home since I use Syncthing on mutiple devices 😅 and every kind of "big data" like films/games I don't feel I need a backup for.
"BackInTime" works similar to TimeMachine, though is somewhat more mighty and complex ... if you want it to be. I fiddled around with rsync + hardlinks, too, but rsync always needs thorough testing. No GUI made it easier, so far. ZFS + snapshots (as btrfs and APFS also do it) is a way to go, but ... who uses ZFS + its sync mechanism on a desktop? The whole paradigm may shift to either mirroring local homedirs and projects to some NAS oder work directly on a share, leaving snapshots and backups to the server, while the OS of the desktop can be restored using some ansible/puppet/....-config onto a vanilla installation.
- 0:00 🚀 Using Linux as a daily driver for 7 years. - 0:11 😅 Lack of a proper backup system despite available tools. - 0:19 🛠 Linux backup tools are complex compared to Apple's Time Machine. - 0:46 🔍 Discovery of Pika Backup, a simple GTK application for Linux backups. - 1:02 💻 Installing Pika Backup on Arch using paru. - 1:16 ⚙ Using version 0.7 of Pika Backup, released days before recording. - 1:23 🆕 Setting up a new backup configuration on an external SSD. - 2:16 ⚠ Warning about formatting the SSD to prevent backup corruption. - 2:55 📁 Formatting the drive to ext4 for use with Linux. - 3:28 🔐 Option to enable encryption for data backups. - 3:46 🗂 Pika creates a self-contained backup repository using Borg. - 4:44 🏠 Default backup location is the user's home directory. - 5:01 ❌ Ability to exclude files or directories from the backup. - 5:42 🧩 Advanced exclusion options using patterns like regex. - 6:24 🔄 Starting the first manual backup with Pika. - 7:03 🔄 Easy restoration of files through the Pika interface. - 7:56 📆 Scheduling automatic backups with various frequency options. - 9:40 🌐 Support for remote repositories via SSH and SMB. - 10:42 💾 Adding existing repositories to new machines for easy data transfer. - 11:19 🛠 Advanced features like data integrity checks and pre/post backup scripts.
Besides not having encryption, timeshift using either rsync (which is what I use) or btrfs seems infinitely simpler. Especially when the entire system breaks and you have to revert a backup, not having to drag and drop entire filesystems, and hope it was fixed. Especially using arch, when a not properly updated package can break the entire system you don't make it super clear how you'd preform an entire system restore. I borked my entire system like 4 days ago because glibc didn't update properly and the entire system refused to boot and brought me to an emergency tty every time I booted, but booting from a live manjaro install (for gui simplicity) allowed easy timeshift restore. The only two things timeshift can't seem to do that this program can is remote backups (but I have 3 free sata ports on my pc aside from the filled ones so I personally don't really care) and encryption, timeshift just seems better.
This (and duplicity) and timeshift have different goals. This (and duplicity) are for data backup, typically /home. Timeshift is for system backup. While timeshift can be configured for only data backup if you want, this pika (and duplicity) cannot do system backup and restore
@@Liriq Then I'd (personally) still prefer timeshift, if there's a specific file you need but for some reason deleted you can open the timeshift snapshots and grab the file. I guess I still don't understand why someone would want something like this and not something that can do most of everything pika can (minus encryption and remote storage, the compression is also pretty nice) over a full system backup. Even using both pika and timeshift in conjunction with each other seems kinda pointless, unless you're regularly switching machines. But even then, if you're doing that I'd want all of my data with me, and as long as you're staying on relatively the same distro (arch based, debian based, etc) timeshift would still handle that fine, with the occasional broken package from switching PM repos. I guess I really don't understand the point of a partial backup system.
@@littlek3000 yup, I prefer timeshift too. Also more sophisticated. Sucks that it doesn't have the option for encrypted backup, which pika and duplicity do
@@b43xoit That's fair, though I've never had any problems with losing files with timeshift/rsync. And it's not like any other backup programs won't also have problems. But if that's ever a concern though I could do a deep dive into how rsync works (I have a lot of exclusions in timeshift mostly because timeshift itself has a TON and because I have a massive steam library, and a 550GB vm drive alone that I don't want to kill a 1TB backup drive with, and I'm not entirely sure how to do exclusions in rsync) and exclude the files I don't want (steam library and VM drive along with the insane amount of already included timeshift exclusions) and run the --checksum flag, which will make it take forever, but *will* ensure all files are synced properly. But I've never had restore problems, and it's way too late and I'm too tired to try to figure this out tonight. Kinda wish timeshift had a built in checksum feature (maybe it does, I searched it up and got both it does and it doesn't answers, maybe it does for set backups, ie, weekly, or monthly, but who knows). If I needed to use the checksum flag, I can only imagine how long the command will be, honestly probably longer than this entire essay of a reply.
Thank you! I'll do a video on my setup one of these days, but basically I use Davinci resolve for most of my editing workflow, with After Effects in order to produce animations.
Great video as usual, but what data do you need backups for? I usually try to keep my computer as data-less (lol) as possible and store all important photos/videos on cloud storages, and keep a setup "script" that I use when reinstalling my system. I suppose storing everything locally and having a backup drive is safer security-wise, but I live in Russia and have some history of political activism (and a mild case of paranoia, I suppose), so I'd like to avoid having my disks seized indefinitely during some criminal case if I ever get prosecuted for saying something illegal in a random comment section...
Video is the biggest one for me. I'd blow through any cloud storage reasonably quick! I use a NAS primarily and also get the benefit of 10Gbe. Any coding stuff I've been fine with a remote git repo! Although it's still nice to have a backup just incase! Yeah, I can imagine it would be nicer to have a little less lying around if that was ever a possibility. Fortunately my content would not cause that to happen 😅
I use Borg as one of my solutions for client's backup since a long time. I did bash wrappers script to handle server backups, using date and time concatenated with server name for each backup name. The script writes states into rsyslog (start / stop / errors / stats) and keeps the whole logging into files that are named the same as the backup, sending email for status / errors / etc. Everything is backed outside the client sites, hosted centrally on my servers and each client backups (which are dedup & encrypted) are even "airgapped" into inaccessible directories on my servers (in case of ransomware, the hackers can't delete all backups!). Borg is very powerful and i use CLI to easily remotely mount backups an restore whatever client ask me! So this Pika solution is a great complement to all this! Kudos for this video!
I've been using Deja Dup for quite some time now. It's pretty similar, at least in the interface and most options. The main difference I see here is that Deja Dup doesn't support patterns that I know of so doesn't have built-in option to ignore caches either.
I'm using Arch, btw :) On my filming machine I have it set up for use with Gnome on X11 as it's easiest to run with ffmpeg, and I have a few scripts to center my windows etc. For wayland, I'll have to use a capture card in order to achieve the same thing.
What is still missing (on Linux) is the way macOS can use the file system event daemon to simply get a list of the files that it needs to backup. Via this fseventd backups take just some seconds, because there is no need to travel through the complete file systems to find altered files, macOS can simply take the list and start copying. A big plus is that timemachine can therefor run every hour without much problems. The Linux pendant could be something like inotify with a suitable daemon in play, but I don't think there is any backup software using inotify.
Or use a filesystem that can make snapshots and send them to a different drive/machine (like ZFS or BTRFS). That's even more fine grained, as the snapshot will only include the changed _blocks_.
@@stephanweinberger That would not include the easy way to make multiple external backups. You can easily define as many backup disk as you like and if a disk is too long away, macOS also needs a full filesystem scan. But it's still nice and easy, and the backup also uses hard links on the backup drives to reduce the amount of space wasted. Let's put it this way: the macOS backup (aka time machine) is one of the best things macOS has.
been running rsync/rsnapshot for like ~ 20 years on server. dropbox and syncthing for desktop/workstation. plus git for projects. never needed much of a desktop backup solution. but i recently started setting up linux desktops for friends & family TM. pika looks like the right tool there. thanks for creating great content!
Right with you, after years of half-assed backup strategies using DAT and CD/DVD-RW, I setup a dedicated rsnapshot box on an SFF system. Dedicated because I realized that I was the primary danger to anything mounted on my primary machines. Quarterly backups to Nov 5, 2004!
Admittedly, rsnapshot is not so user-friendly as this tool, but in exchange I can use Linux knowledge to accomplish useful tricks, like moving to bigger drives, mirroring to an offsite archive, and fixing things when something goes awry.
I'm curious, do you encrypt your rsync/rsnapshot backups? I usually make a tar.gz file, encrypt it with gpg and then scp to a server.
It has the advantage of being seamless (gpg asymmetric encryption does not require a input), but that does not make incremental backups
@@nogesax3535 In my case, I have a local rsnapshot server, 2 external mirrors (rotated off-site once a month), and a Google Compute Engine mirror. All backups are on encrypted filesystems (via LUKS). Encrypting at the file level would be somewhat opposite of how rsnapshot works - not saying it's not possible, just that I worry that it would be hard to manage successfully and reliably while also not leaking data.
in my case no encryption. external hard drive. same disk for almost the full time. replaced once to upgrade size. old WD "Book". so incremental works ofc. backups usually take only seconds.
Hear... hear... #metoo.
Git is the BEST incremental backup tool ever!
And yes, non-incremental backup is done by rsync.
This is one of the very best linux software tutorials I have ever seen. Very detailed but NO fluff and unnecessary "background music". Well done. earned a like and a sub
What do you have against background music?
Anyone who has hearing loss has a lot of difficulty separating the music and the narrative. That is my rationale for disliking "background music". The second reason is that some folks do not know how to level the audio narrative so that it is consistent across the records so you constantly have to tweak the volume control up or down to be able to understand the narrative. Thanks for asking!
@@wesbryie
same here!
@@wesbryie Most of it is GAY...
I was using Pika before watching this video for personal file backups and Timeshift for system snapshots
I wish they'd join hands and come up with a unified intuitive interface for both types of backup.
I used Deja Dub instead of Pika but I will try Pika. Deja Dub was just the recommendation in Gnome Software
@nico1337 where do you keep your system snapshot? On an external or separate partition?
@@tennisfreak312 on one of my 7 internal drives
@@tennisfreak312 Same problem that I had when it came to backup-ing, that I switched back to Windows after almost breaking my computer while trying to restore my system... 😅
Not only the topic is interesting. I'd like to acknowledge separately the level of video production with those blur and highlight effects. Does feel classy comparing to our usual YT Linux stuff videos (not that those are necessarily bad, but...). Geez, an instant subscription.
in kde systems there is KUP, which makes incremental backups and is integrated with the system itself. The behaviour of KUP is similar to file history in windows, it takes backup on the specified interval and auto detect if the specified drive is mounted or not, if not then it will not do anything and as soon as you connect your drive it automatically take a backup. the ease of use of KUP is identical to file history in windows. You can access kup in system settings in kde and setup multiple schedules if you like
It also supports making a full backup every time, rather than an incremental backup.
@@contrabombarde1216 true, forgot about that, and that is also quite intuitively available
Doesn't seem to be there in Plasma 6
@@timotiuswidjaja2792 Still on Plasma 5.27, but I really hope they didn't remove it. That would really suck.
quick tip to pika devs: if you're targeting user friendliness for the interface, maybe don't say 'repository' and just say backup location. not everyone is a developer and knows what 'repository' means. this could cause some mild confusion when they click backup but are then confronted with a question about setting up a repository and going, "huh? wait what's a repository? I just want to create a 'BACKUP' "
100%. I'm a completely new Linux user. Been using it for one week at most. Prior to this, my experience with GitHub amounted to knowing its name. I had also never heard the term 'repository' before, and I'm still somewhat confused on what it means.
Another good Borg backup client is Vorta. You can exclude certain folders by making a .nobackup file under the folder.
Love the Star Trek names ❤
vorta is a python app. pika backup is a rust app, that means it takes a long time to compile using AUR in Arch Linux :(
in my test VM i enabled chaotic AUR for fast install aur precompiled stuff
I have nothing setup at the moment and this is exactly what I was looking for. A life saver I must say.
I also love the open nature of the tool, that it has a simple side and also a full on complex side.
How the hell do u have so few subscribers, This is some really high quality content. I really look forward to more of your vids!!
Thank you!
I think it is because, the first Video is only a Month ago. It needs a bit of time to gain subscribers. But I think for the short time, that this Channel is active, it gains a much of subscribers. But this Channel needs more subscribers. I already love this Channel
@@dreamsofautonomycan u also use the software to restore folders to there diffrent respective locations?
He is also the person behind this channel
youtube.com/@dreamsofcode
Let's say it's not his first rodeo ahahah
Keep up the good job
Because this is his second channel
*It's a real shame you can't find this great video if you google for Pika Backup* - or only if you're lucky. A title like this would be much better: *Pika Backup: We finally have a simple solution for backing up files on Linux*
Hey just to let you know that Deja Dup Backup also existed for a long time and similar to this but this tool looks very awesome and probably way better/polished and advanced.
I use Deja Dup and it is easy to setup with automatic backups. It is just GUI for tool called duplicity.
@@tom3f yeah I know that one, it comes with Manjaro for example, it's not bad 😁
I'm using Deja Dup..no issues so far and very simple to setup.
Great stuff!
I'm a simple user, so this might be too simple, but I only use `rsync` and a bare git repository to backup my entire machine. A post-installation script will restore all of that along with installing packages in 30 minutes, I can reinstall my machine right now without the fear of losing anything.
Scheduled script every hour with rsync incremental mode here. Works easy and super fast .
I've been a happy Back In Time user for about a decade. It's rsync-based and a good balance of simplicity and features.
I read an allegation (sorry source amnesia) that `rsync` contains an error and can corrupt files.
@@b43xoitnever happened to me and my years worth of backing up and restoring stuff....
@@b43xoit Considering how much rsync is used in different contexts, that sounds a bit strange.🙂
Excellent tool and presentation.
I used Bjorg, but not with so much settings. This is far easier with Pika to achieve a full featured backup flow.
For the longest time I was just stuck with using a mess of an rsync shell script to have some resemblence of version control and such. Been tempted to get into borg but the barrier of entry felt very daunting. Pika on the other hand looks amazing, need to try it out ❤
Let's be honest, the user experience for Pika (like a lot of gnome) is very nice indeed - Linux has always been powerful, but now it's starting to get the whole design/ux piece right. As a recent convert, I've been messing about with snapper and btrfs assistant - I'll have a look at Pika :) Awesome video, thank you!
Honestly, I'm really vibing with what Gnome is doing at the moment. The level of consistency and ease of use I think is going to push the platform higher when it comes to mass market. It's almost at the stage where I could start to recommend it to non technical people.
I am using Vorta and it served me well - but pika looks a bit more polished for "simple setup and every day usage" for people who don't want to understand the inner workings of borg. I am happy that you covered it - I will just send anyone I find without a borg backup to watch this video in the future
Backed up Garuda Linux to my Synology NAS without issues. Thanks for the tutorial 😊
At present, simple does it. Straight forward bash scripts to rsync to a couple of ZFS repositories with snapshots before and after. It may not be the most elegent, but easy to understand, and has got me out of trouble more than a few times.
I would not install or use z. FS on an envy M. ESS d until they fix the piss pour. Random reed and writes performance out of the box by default
From what I understand there isn't simple file. Or command that 1 must edit. I think involving something about highsync or usync =13 ...to fix it?
I'll just stick to the tried and true ext4
@@motoryzen People don't use ZFS for top performance, they use it for data integrity and/or its management features. For what it's worth, ZFS works fine on my NVMe machines (not "piss poor" as you might say), but of course it'll always be slower than ext4 or XFS simply due to the extra checks that ZFS does.
I read an allegation (sorry source amnesia) that `rsync` contains an error and can corrupt files.
@@austinleong3319 yes I know what they use it for but the bottom line is the random reading right performance is so shity compared to the normal standards that we've been used to with nvme technology that it basically defeats the point
One and even better analogy to see what I mean go back to dial-up internet
Auf den Punkt. Kein Stottern. Keine "Ääähs". Keine "Eeers". Einwandfrei gemachter Beitrag, mit dem ich meine Zeit nicht verschwendet habe. Von der Art Vorstellung will ich mehr sehen. You are subcribed! :)
Thank you muchly for all the learnings. Wishing your channel much success!
Used mountable sector-wise backups of the whole partitions: zerofree+dd+mksquashfs with pseudofiles compressing on the fly.
Used zpaq for multiversioning of one directory. Not especially user-friendly but achieves maximum compression+multiversioning+deduplication.
timeshift does just fine as a backup solution and works like time machine, only with more user configurable options. It's meant just for system backups, but you can include user files as well (or just your dotfiles). Backintime is also pretty good. Fully featured and nice GUI. I don't understand how this fails the simplicity test. Configuring timeshift takes about 10 seconds with the GUI client. Not saying Pika is in any way bad, but Linux has not been short of backup solutions.
Not a lot of mentions (if any) for restic, it’s not as simple as Pika, or rsync but I’ve found it fantastic for backing to object storage, it does it using packed blocks and has a repository file saving non-negligible amounts of money with transaction costs especially when backing up source code.
I took a look at Borg a while back and couldn’t get my head around it for my use case (full disclosure I may have not tried hard enough).
But yeah restic has been awesome, backing up multiple hosts efficiently and recovery has been straightforward.
Worth a look I think.
Oh just to add, it has a sparse checksum validation feature as well to do integrity validation on random subsets… big win when you have a large backup set.
Great vid! Looking forward to the borg vid!
Me too! Borg is really dope
NOOOOOOOOOOOOOOOOOOOOO!!!!!!!!!!! They assimilated my people!!!
@@dreamsofautonomyI can remove most of your implants and turn you back into a human! RESISTANCE IS NOT FUTILE!!!
@@dreamsofautonomyisn’t Pika essentially a gui frontend for borg?
@@Batwam0YES like Vorta is a gui for borg, too
Just using rsync. Great video. Thanks
Using restic for my homelab servers. One local hard drive repo receives the files from all the servers. The repo is then synced off-site with rclone. All that process is automated via a few basic scripts.
Pika looks very nice for a desktop use case! I wish there were the same UI for restic so I could add a desktop backup to the same repo I already use (though I have limited need for desktop backup, using a dotfile git and storing files on the servers rather than locally)
i have very similar protocol, but i use restic copy snapshot instead of rclone. is rclone faster than copying restic snapshots?
@@bernardcrnkovic3769not them, but not sure, there is also restic-rest server. I suppose in theory it's supposed to be faster. It would mean going through less protocols, too. That is where some inefficiencies can happen with chunking between protocols. I'd be curious in what people find though
Very nice job with explaining Pika Backup. Thank you.
Thanks, I was looking for backup solutions for a new Linux installation. One thing though, you don't need to use the time machine interface to browse files, you can just open the backup drive and access one of the time stamped backup sets.
1:08 "I'm using Arch btw" that was a cheeky flex
Arch is a noob/meme distros nowadays.
ur a meme
Nice, good improvement over other options. That said, it's way more fun to setup another system as a server, install a RAID drive array, then mount it as an SSHFS mount, then write an rsync script to backup everything I want, then run the script via Cron job. I don't do things the easy way :-)
i use a script that uses borg to backup my markdown notes to my USB stick sitting on my fritz router mounted as cifs
that way if my hard drive dies then at least my notes are in a more current backup state than the real backups i don't do that often to external USB drives
Nice to see the channel is still going.
Hope is grows.
Thank you for making this. This is so much better and easier to use.
Pika Backup on Linux-mint and presumably Ubuntu, 760Mb download, 2.6Gb on disk (Flatpak). Now, I don't know a lot about computers but that sounds mindblowingly bloated to me for something that I could fashion with rsync and a bash script in just a few dozen bytes.
Wow. What a great solution. Excited to use this. It looks soooo easy.
Was thinking of Borg right from the start, having used it for the past 6-7 years, and lo it turns out to be a borg GUI! Pika looks like a pretty good tool to get to an initial borg config even for the cli-inclined.
Clean video production! I like Pika Backup, but since full system snapshots aren't really a thing on Linux (other than Timeshift and btrfs assistant which required arduous tinkering to set up), I just back up my entire Home folder in Dropbox in an encrypted .zip file, along with a post-install script that installs everything on a fresh install.
For so many months I have been facing same problem of not having a simple backup system thanks bro love your every video.❤️
Pika sounds promising. I've had a good experience with duplicati on desktops, and tar incremental backups on servers.
I use my own shell script run as a cron job.
It backs up to a selection of USB thumb drives. For each type of backup (home, /usr/local, mail, /etc) it:
* deletes all but the latest N (usually 2) copies by date from all backup drives,
* creates the backup in a temporary space,
* copies it to multiple thumb drives (usually 3).
I’m glad I bumped into this! Now all I wish to try is, hope it can backup my whole distro and cache files incase if I wish to to move to another distro like popOs to fedora! Gonna check this out. Have gladly subscribed ! Thanks for the video 🎉
I use Rsync, is very good. Excellent videos, have learned a lot.
I read an allegation (sorry source amnesia) that it contains an error and can corrupt files.
I've been using syncthing while I was developing my own solution for backups. After seeing this and giving Pika a shot, I'm happy to say that I will no longer be hearing my laptop fans spin up and my computer lag many times a day as syncthing does its syncing.
I'm using Synology Drive. I don't backup whole system, just documents and project files - source code, images etc. It works really well for me. For backing up applications, I created sh file which will install all apps (flatpak and deb packages) on new machine. I don't backup whole system, because I build large projects - linux kernel, PrusaSlicer and other SW, so build deps would take a lot of space for backup.
This is a wonderful, and timely video, (for me). I've been looking for something like this since I first started using Linux years ago. I'll be looking forward to setting this up on my system and experimenting with it. It really looks promising. Thanks for making this video. I'll be looking forward to your future videos on the subject.
/Bob (A Now-New Subscriber)
Super helpful, exactly what I needed while trying this myself. Thank you
this is awesome, looks great! We always had a simple solution for backing up files on Linux with rsync though
i use rsync and can testify its more complex than any windows backup tool i've used
I read an allegation (sorry source amnesia) that `rsync` contains an error and can corrupt files.
@@b43xoitAnd that "error" would never have been solved, while rsync is as old as Linux is and is the basis of about any back up system out there? Hard to believe.
Pair this with Timeshift and you will have a complete backup system.
Timeshift (or snapper) combined with bootable snapshots makes my ArchBTW feel like DebianFYI.
Great tip! I'm gonna nerd out on Timeshift later tonight.
@@snickersanyoneTou spelt DebianFTW wrong 😊
Yeah, I also went with Timeshift for the System and PIka for the personal files. Timeshift saved my ass every-time I tried something crazy enough to destroy my whole system.
Timeshift has saved my butt a few times..... usually mucking around with new GPU drivers.
SO... this is a GUI frontend for the actual and awesome borg-backup
This channel is an absolute goldmine wow
Which terminal do you use? I love the aesthetic.
I'm loving your videos, by the way. Solid content!
That's Console it's part of Gnome. But gnome has 2 terminal applications one is Gnome terminal and the other is Console
@VitisCZ I'm actually using Gnome Console, but it doesn't look that. Is that a theme or something that you have to configure?
@@TheALPHA1550that’s Zsh with the powerlevel10k theme
He uses Alacritty.
I used Vorta till today. Indeed Pika is friendly and easier.
I use rclone with pcloud and icedrive. Will give pika a spin. Thanks for the informative content.
Pika Backup is a frontend for Borg Backup. As used with the CLI you also can use Borg directly. Not hard at all, has a lot common with git. Instead of Pika Backup i now used "Vorta Backup" instead. Vorta also is a frontend to Borg Backup. I picked Vorta because it already ships as a Debian default package, so no flatpack needed. But Borg itself is really easy to use, so in general great option!
Used Deja Dup before but it somehow is slow in backup and restore. Also one time my backup was broken and it just told me that everything is lost and I need to rebackup from start. Usually not what you expect from a Backup program. A Backup program should be usuable even if the backup itself is broken. At least Borg seems to provides a check, repair and also provides option to use it even with broken files.
Its also extremly fast. Good for Desktop backup. For more enterprise stuff "Bacula" is a very good option. Used it before and replaced it by horrible rsnapshot.
Honestly one of the reasons I love openSUSE is how easy snapshots are with it. Yes I know, snapshots are not backups and I do have my important stuff secured elsewhere but for the most prevalent problems: Broken updates or breaking things with configuration, it's enough.
my system backups were living on a prayer.
if I broke my system, I would have to manually plug in a live USB with linux, mount my laptops drive and partitions, and search for the files manually (hoping they are still their).
Now that I know of this tool, I have already started downloading it while watching the video :)
Thank you so much for a very good review. Very helpful, indeed! I'd be grateful if you can top this off with a video on Borg backup some time in the future!!! Keep up the good work, your channel is amazing!
I wasn't sold on it until. You showed Pika can browse the archive. Yes please.
Every complaint i had with linux backups solved by this one application. Finally!
TimeMachine is one of the most important reasons I still use OSX/MacOS. It has saved me twice, once restoring from to a completely replaced drive.
Not just my files but my full OS including all settings for printers, wifi and everything else.
It is simply brilliant, I really don’t understand why Windows and Linux don’t have something similar. It is from 2007, so it completely baffles me that no other OS has been able to create something similar in 17 years..
OS-X puts Linux to shame. Long time user of both platforms
as a regular non-tech-savvy user, when it comes to backing up personal files, it's rather simple and easy. I know which folders to backup. I was wondering more about what should be backed up for reverting a system back to it's previous state after a system update that went wrong, or something like that. But I don't even know whether that's possible to do on a mutable system. I guess I should search for some tutorials about it.
If you just backup your entire root directory you should just be able to copy it back and be fine, at least if it's all on one physical device. If you have a weird partition structure with other drives mounted on your fs then you might need to recreate that manually before you restore. The only issue is your /boot volume, whether it has the right kernel image for the hardware you're running on (if you need to switch to another machine) and the boot loader / uefi partition. But if you run a linux install disc with minimum options and then just copy your files back that would probably be fine. Just exclude /boot from the backup.
Now you can make it harder if you want to actually save space and not back up things like package files, but that can get complicated. If you're using an arch based flavor I can tell you that /var/lib/pacman stores metadata about all its packages, including file hashes. But you need to write/find a C program that calls libarchive (with gzip and mtree extensions) to read it.
Or you could just exclude the output of "pacman -Ql" or whatever command lists all files on your system that comes from packages (apt list?). But then you'll lose anything that you modified, like /etc/mirrorlist
Thus channel is a gem💎
Please keep up with quality content
I use BTRFS snapshots for incremental copies and dd's for full disk images. Scheduled Rsync jobs also work.
This is one of the best videos that I've seen recently, thanks for sharing
I'd just like to simply copy data - no repository BS, no "restore later" nonsense, no compression. Just copying over and check for duplicates (and overwrite them). I want to backup my data for use later, not to bind them forever onto Linux/UNIX. This is a nice tool for a crappy backup logic behind (borg).
I literally just came back to Linux on my desktop and was in need of a backup solution. Perfect timing
thanks! needed just this video.
1:56 what machine do you use? seems compact and probably much quieter than the conventional tower I have
Looks similar to a Beelink minipc
As an openSUSE Leap, KDE. user for years now I will see if I can apply what this Excellent and helpful video has shown me, to my system, using a spare Crucial MX500 sata ssd. Even having some knowledge of the CLI, the prospect of Linux backups has, until now, seemed off putting . Great work. Thank you and subscribing.
This video is done so nice and your voice is so nice. Instant sub! Keep up this great work.
Oh nice, I have been using duplicity mostly, but this looks really great!
If you cannot select a certain filesystem for formatting, that means you do not have that filesystem installed (such as ntfsprogs or exfatprogs). for example, fat and ntfs are often installed by default, but f2fs is less often installed by default. There are options under Other that may be available, it just doesn't show them on the first backup page.
I'm looking to implement a strategy where I sync all my persistent data (documents, music, etc.) to my NAS via Syncthing and then just do the backups server-side. Since I use NixOS, any system configuration can easily be reproduced with a few commands. I'd only really miss out on backing up "temporary" data like local container volumes, browser cookies, downloads and maybe some saves from Steam games.
Agreed! I've been playing more and more on nix on my main workstation and I think Pika/Borg goes nicely together. You can also manage Borgmatic from your nix configuration as well.
Since I'm on butter I use Timeshift to snapshot my filesystem hourly. Since I don't really have anything important on it, I don't bother backing it up externally. Everything of importance is on my NAS and modern SSDs are reliable enough for me to not bother as much
Thank you for this. I've been looking for a BU app that was easy (so that I'll use it) and also writes accessible files to the archive. Having been on Retrospect for many years, I was always concerned that the archives were not even mountable on any OS, let alone retrievable. It seems the only thing missing from Pika is the last item on my list: I wish it was multi platform, or at least on Linux and Windows. Oh well.
I use a combination of Syncthing and Duplicati. Syncthing is very useful for network drives or remote PCs. Then I use Duplicati which is installed on a docker container on my Pi5 to do encrypted backups to my NAS that is running in raid 1 and another in raid 6. 😌
What micro computer are you using? Would you recommend? I’m in the market for something besides a laptop. I don’t play games, I’ll run a nice linux wm, and do development. Thanks.
Hello.
Thanks for the video.
I was using rsync for years and now i use timeshift on my fedora 40
I have a Raid of 2 8TB drives with all movies, music & photos backed-up, and all family members can get to them.
I have another Raid of 2 8TB drives, and use a shell script I wrote to backup weekly.
Thanks for the great video! I have one question, if you don’t mind. Would this be a good solution for a frequently detached external drive? I’m planning to use Picka to back up my personal laptop every weekend, for example, only when I connect the laptop to the external drive.
Looks good, but I've been searching for incremental backups. We used to do a Full one week, three Incremental weeks, and keep the last two months. This system was trade off between space and finding files. You only had to look through a few Incrementals to make sure you had the latest file.
Great quality, informative video. Liked it very much!
I had some issues with Pika that made me switch away. It would only work in Gnome, so I would be unable to access my data id using a different desktop. I would also get issues after mounting a lot of backups in one session, so I would need to close the app and reopen it.
I switched to Vorta instead, which didn't have those issues. It even shares the same backend, so my Pika backups still worked with Vorta.
I don't really need a backup solution for home since I use Syncthing on mutiple devices 😅 and every kind of "big data" like films/games I don't feel I need a backup for.
"BackInTime" works similar to TimeMachine, though is somewhat more mighty and complex ... if you want it to be.
I fiddled around with rsync + hardlinks, too, but rsync always needs thorough testing. No GUI made it easier, so far.
ZFS + snapshots (as btrfs and APFS also do it) is a way to go, but ... who uses ZFS + its sync mechanism on a desktop?
The whole paradigm may shift to either mirroring local homedirs and projects to some NAS oder work directly on a share, leaving snapshots and backups to the server, while the OS of the desktop can be restored using some ansible/puppet/....-config onto a vanilla installation.
- 0:00 🚀 Using Linux as a daily driver for 7 years.
- 0:11 😅 Lack of a proper backup system despite available tools.
- 0:19 🛠 Linux backup tools are complex compared to Apple's Time Machine.
- 0:46 🔍 Discovery of Pika Backup, a simple GTK application for Linux backups.
- 1:02 💻 Installing Pika Backup on Arch using paru.
- 1:16 ⚙ Using version 0.7 of Pika Backup, released days before recording.
- 1:23 🆕 Setting up a new backup configuration on an external SSD.
- 2:16 ⚠ Warning about formatting the SSD to prevent backup corruption.
- 2:55 📁 Formatting the drive to ext4 for use with Linux.
- 3:28 🔐 Option to enable encryption for data backups.
- 3:46 🗂 Pika creates a self-contained backup repository using Borg.
- 4:44 🏠 Default backup location is the user's home directory.
- 5:01 ❌ Ability to exclude files or directories from the backup.
- 5:42 🧩 Advanced exclusion options using patterns like regex.
- 6:24 🔄 Starting the first manual backup with Pika.
- 7:03 🔄 Easy restoration of files through the Pika interface.
- 7:56 📆 Scheduling automatic backups with various frequency options.
- 9:40 🌐 Support for remote repositories via SSH and SMB.
- 10:42 💾 Adding existing repositories to new machines for easy data transfer.
- 11:19 🛠 Advanced features like data integrity checks and pre/post backup scripts.
Besides not having encryption, timeshift using either rsync (which is what I use) or btrfs seems infinitely simpler. Especially when the entire system breaks and you have to revert a backup, not having to drag and drop entire filesystems, and hope it was fixed. Especially using arch, when a not properly updated package can break the entire system you don't make it super clear how you'd preform an entire system restore. I borked my entire system like 4 days ago because glibc didn't update properly and the entire system refused to boot and brought me to an emergency tty every time I booted, but booting from a live manjaro install (for gui simplicity) allowed easy timeshift restore. The only two things timeshift can't seem to do that this program can is remote backups (but I have 3 free sata ports on my pc aside from the filled ones so I personally don't really care) and encryption, timeshift just seems better.
This (and duplicity) and timeshift have different goals. This (and duplicity) are for data backup, typically /home. Timeshift is for system backup. While timeshift can be configured for only data backup if you want, this pika (and duplicity) cannot do system backup and restore
@@Liriq Then I'd (personally) still prefer timeshift, if there's a specific file you need but for some reason deleted you can open the timeshift snapshots and grab the file. I guess I still don't understand why someone would want something like this and not something that can do most of everything pika can (minus encryption and remote storage, the compression is also pretty nice) over a full system backup. Even using both pika and timeshift in conjunction with each other seems kinda pointless, unless you're regularly switching machines. But even then, if you're doing that I'd want all of my data with me, and as long as you're staying on relatively the same distro (arch based, debian based, etc) timeshift would still handle that fine, with the occasional broken package from switching PM repos. I guess I really don't understand the point of a partial backup system.
@@littlek3000 yup, I prefer timeshift too. Also more sophisticated. Sucks that it doesn't have the option for encrypted backup, which pika and duplicity do
I read an allegation (sorry source amnesia) that `rsync` contains an error and can corrupt files.
@@b43xoit That's fair, though I've never had any problems with losing files with timeshift/rsync. And it's not like any other backup programs won't also have problems. But if that's ever a concern though I could do a deep dive into how rsync works (I have a lot of exclusions in timeshift mostly because timeshift itself has a TON and because I have a massive steam library, and a 550GB vm drive alone that I don't want to kill a 1TB backup drive with, and I'm not entirely sure how to do exclusions in rsync) and exclude the files I don't want (steam library and VM drive along with the insane amount of already included timeshift exclusions) and run the --checksum flag, which will make it take forever, but *will* ensure all files are synced properly. But I've never had restore problems, and it's way too late and I'm too tired to try to figure this out tonight. Kinda wish timeshift had a built in checksum feature (maybe it does, I searched it up and got both it does and it doesn't answers, maybe it does for set backups, ie, weekly, or monthly, but who knows). If I needed to use the checksum flag, I can only imagine how long the command will be, honestly probably longer than this entire essay of a reply.
Excellent video!
What do you use to do your videos editings?
Thank you! I'll do a video on my setup one of these days, but basically I use Davinci resolve for most of my editing workflow, with After Effects in order to produce animations.
Timeshift, Deja dup and Lucky backup is both simple, but generally U R right.
Great video as usual, but what data do you need backups for? I usually try to keep my computer as data-less (lol) as possible and store all important photos/videos on cloud storages, and keep a setup "script" that I use when reinstalling my system.
I suppose storing everything locally and having a backup drive is safer security-wise, but I live in Russia and have some history of political activism (and a mild case of paranoia, I suppose), so I'd like to avoid having my disks seized indefinitely during some criminal case if I ever get prosecuted for saying something illegal in a random comment section...
Video is the biggest one for me. I'd blow through any cloud storage reasonably quick! I use a NAS primarily and also get the benefit of 10Gbe.
Any coding stuff I've been fine with a remote git repo! Although it's still nice to have a backup just incase!
Yeah, I can imagine it would be nicer to have a little less lying around if that was ever a possibility. Fortunately my content would not cause that to happen 😅
Vorta too uses Borg and I much prefer it - both it and Pika are excellent tho with intuitive GUI and sane defaults making it very user friendly.
I use Borg as one of my solutions for client's backup since a long time. I did bash wrappers script to handle server backups, using date and time concatenated with server name for each backup name. The script writes states into rsyslog (start / stop / errors / stats) and keeps the whole logging into files that are named the same as the backup, sending email for status / errors / etc. Everything is backed outside the client sites, hosted centrally on my servers and each client backups (which are dedup & encrypted) are even "airgapped" into inaccessible directories on my servers (in case of ransomware, the hackers can't delete all backups!). Borg is very powerful and i use CLI to easily remotely mount backups an restore whatever client ask me! So this Pika solution is a great complement to all this! Kudos for this video!
Would you be willing to share more information and/or instructions on how to do what you do? Maybe a DIY video? Thanks in advance
@@jim7smithOk, i'll check the way i'll share this and report back here!
@@guyboisvert66 Thanks so much, looking forward to seeing the info
@@jim7smithWhat's your level as a Linux User / Sysadmin? Experience in bash scripting?
Great video. What software do you use for the smooth zoom in and out for the screen share? It’s like butter
What gnome theme is this? Love the borderless app headings. So clean!
I've been using Deja Dup for quite some time now. It's pretty similar, at least in the interface and most options. The main difference I see here is that Deja Dup doesn't support patterns that I know of so doesn't have built-in option to ignore caches either.
Nice video. One question though, what is your current OS and OS theme ? it's simple and beautiful
I'm using Arch, btw :)
On my filming machine I have it set up for use with Gnome on X11 as it's easiest to run with ffmpeg, and I have a few scripts to center my windows etc.
For wayland, I'll have to use a capture card in order to achieve the same thing.
Great video. Good pacing, good coverage, very clear.
And what a great piece of software! I have to look into that.
What is still missing (on Linux) is the way macOS can use the file system event daemon to simply get a list of the files that it needs to backup. Via this fseventd backups take just some seconds, because there is no need to travel through the complete file systems to find altered files, macOS can simply take the list and start copying. A big plus is that timemachine can therefor run every hour without much problems. The Linux pendant could be something like inotify with a suitable daemon in play, but I don't think there is any backup software using inotify.
Or use a filesystem that can make snapshots and send them to a different drive/machine (like ZFS or BTRFS). That's even more fine grained, as the snapshot will only include the changed _blocks_.
@@stephanweinberger That would not include the easy way to make multiple external backups. You can easily define as many backup disk as you like and if a disk is too long away, macOS also needs a full filesystem scan. But it's still nice and easy, and the backup also uses hard links on the backup drives to reduce the amount of space wasted. Let's put it this way: the macOS backup (aka time machine) is one of the best things macOS has.
Vorta is a good Borg frontend if you use a Qt based DE.
lol. Do you think my mom would also use paru on arch? Or would she just buy a mac and have it work without doing anything?
Man, what desktop environment are you using in the video? - it looks awesome.
In this video is pure gnome! Although I use hyprland these days
@@dreamsofautonomy Got it. But did you customize it through Arch options or how did you get that awesome theme design?