Ask HN: What are your peronsal data backup and sync setups?
I am the OP and I see the post didn't see much traction. Anyway here is my setup:
- All my personal data: Borg (gui: vorta), Restic (gui: Backrest)
- Very important personal data (small subset of all data): Tarsnap (gui: tarsnap-gui)
Sync:
- Personal regular use data data: Dropbox (some private/personal data that I want to be extra careful about gets synced in a Cryptomator folder and in a KeePass file), iCloud (not a lot of files)
- Photo/video: iCloud, Ente
----------
I am also trying to setup e2ee backup/sync on my Koofr a/c with rclone (WIP) for almost all my data - more than "personal" data (but not the system files of course)! Koofr a/c is the lifetime subscription and even they they have a great track record I do not really consider it "lifetime". I also hope they expand Vault (i.e not limited to just only "one" folder)
Yeah great question. With family members including not living with you. It's virtually impossible to backup from iPad/Phones to android phones and tablets and multitudes of PC's Linux to iOS to Wintel.
Eagerly awaiting answers from you all but for now...
1. rsync from where I can.
2. manually copy whenever I get the device in hand.
3. backup cronjobs from omv server
4. my main desktop use Timeshift
5. LocalSend
Also cronjobs to attached USB and NAS to have backups of backups of backups.
No one seems to realize every household is a micro enterprise!
Then make matter worse, with all the tightening of security. It's harder and harder to remote into and recv/xfer files l.
> Yeah great question. With family members including not living with you. It's virtually impossible to backup from iPad/Phones to android phones and tablets and multitudes of PC's Linux to iOS to Wintel.
It’s simple with Google Drive or Dropbox
Yes but I'm inclined to not allow "public" cloud of personal photos. I'm almost positive all your files are being used to train AI. Not to mention nothing works on iOS except apps from aapl.
So long term, I'm thinking of private cloud and a cross platform app where its sole purpose in life is to xfer all files in a particular folder on the device to the private cloud.
Most are not tech savvy enough to setup that backup. Not to mention I'll need access to aquire.
Long list of why's.....
Neither is true. Google Drive, Dropbox, OneDrive, and even Amazon Drive work on iOS and all have policies that they don’t train on your photos.
Logically, how could they train on your photos? Training on photos require humans to label the content of photos for training.
All of the apps I mentioned are cross platform.
Have you tried to exchange photos from iOS to Android? Also Google drive requires you to have a Gmail address.
Not well versed in all platforms but whenever I talk to others to exchange photos of events and such. It's always a tremendous pain point to figure out how/where to.
Just my observations as I am trying to figure out the most frictionless way to exchange pics.
Also reminder that each versions of iOS and Android has their own limits based on app level that tops out at their versions.
Just a big huge mess.
PS. imho. Policies don't mean jack when there is absolutely no way for you to know if they have or haven't scraped. Gmail eng already got caught using his level of access to harass ppl. And Ring /Tesla were both caught sharing what should have been private and anon videos. I understand from dev point one must be able to decrypt to debug but nevertheless, you use their service, it's theirs to do whatever they want.
Not that I'm peddling anything illegal but my family photos I've already had random files including pictures and videos that wasn't mine in my Google drive and it also wasn't wrongly shared.
CV+ML with AI now don't need human intervention to id the materials.
Again none of this is true. Google drive doesn’t require you to have a Google address - just an email address when you create an account.
On iOS, once you install Google Drive, it shows up as a storage provider in the Files app - as does One Drive, Dropbox, iCloud Drive etc abd you can copy files back and forth. You can also connect any USB C storage device to your iOS device and it also shows up in the Files app and you copy files to it.
AI is nothing magical. What good is using random photos for training without labeling?
> Also reminder that each versions of iOS and Android has their own limits based on app level that tops out at their versions.
What does that even mean?
And Tesla is not exactly a trustworthy company nor is it’s leader
My observations are from doing the grandma test. Just trying to figure out how to reduce or eliminate the pain points I've observed.
I have a "stuff" git repo where each machine has a branch. Anything I create that should persist more than a day or two but doesn't have another home gets committed to that and pushed to a VPS.
I've been experimenting with different systems for archiving binary files like images but so far there's nothing I've stuck with. I just don't keep that many of them so the few that I do go in ~/stuff and just consume twice the disk space they should.
One nice thing is that I have bash keep its history there with datestamps and hostnames. So c-r works on all my machines (I typically merge every time I switch between computers) which is super cool.
Main machines are zfs, with sanoid to take snapshots and syncoid to transfer snapshots among them.
Windows machines backup nightly with Veeam personal (free) license to one of the servers. I didn't have luck with open source windows backup.
That's also how I do it. Great setup, understanding ZFS is totally worth it.
On non-zfs+non-windows Clients (MacBook and Linux Notebooks) I use a restic cronjob.
Used to be restic[1], but I switched to kopia[2]. For Android I just sync /storage/emulated/0/ with syncthing. All devices are backed up to a home server. Home server backups go to Backblaze (but any S3-compatible storage would do).
For important files I have my own bewCloud instance (Nextcloud alternative I've created) which syncs directories via rclone. For system backups I have physical SSDs that I connect via USB (using Déjà Dup for Linux, Time Machine for macOS). Every month I'll also do a separate encrypted backup of my home directory to Backblaze. For my GrapheneOS phone, I don't care (it's barely got anything). For my iPhone, iCloud.
Simple:
iCloud + Google Drive + OneDrive
Photos and videos are automatically synced to all three. My few important documents are copied between all three manually via the Files app on iOS.
I used BackBlaze in the past when I had a large media library of videos that fell off the back of a truck. But once I took down my Plex server , I copied all of my media to my personal AWS account to S3 Glacier Deep Archive. It’s less than $2 a month for 2 TB.
Every week or so, I plug in an external SSD and put a tarball on it of the directory hierarchy in which I keep my most valuable data. If the tarball is significantly larger than previous one, I use du to try to find what caused the increase and move it out of the hierarchy.
For me, it's easier when I have everything done automatically. Thus, I prefer using backup tools to set up a backup copy to be run automatically. For that I use GitProtect
I sync my devices to my Synology NAS at home, and then the NAS syncs them to Backblaze B2.
I have some extra workflows to back up some other data, like some WebDAV shares, based on rsync.
tar | zstd | gpg > USB
https://gitlab.com/sdwolfz/dotfiles
```sh
make backup
```
Sync between phone and laptop is set up with Syncthing.
rsync and btrfs snaphots internally, and restic for important backups synced to external storage. Cron for scheduling. Trading bits of NAS space with friends for geographic redundancy without needing to pay backblaze/etc.
The Backblaze cloud and restic.
I just use Backblaze these days, now that I'm out of being a sysadmin.
Backblaze and a Windows network share of the backed up folder.
syncthing from desktop and mobile to homelab. rsync from homelab to raspberry pi with a direct attached storage device at my parents.
Syncthing plus Duplicati2.