When all parts come together 😉 I now have S3 compatible storage with #garage in my homelab, using #nginx as reverse proxy and secured with a certificate from my own #StepCA based CA (Certificate Authority) that gets auto-renewed by #certbot. And this all works without any internet connection, as I also have a DNS server for my home network with the correct CNAME entry for s3.
Anyone using S3 (compatible) storage for incremental backups of Linux and/or MacOS machines at home? If yes, what backup software/tool? Open source preferred. My first unfocused research suggests to look at restic [1] and maybe just some scripting around rclone [2]. Any hints/links welcome! (Yes, this is part of my project to use local S3 compatible storage based on garage [3])
[1] https://restic.net
[2] https://rclone.org
[3] https://garagehq.deuxfleurs.fr
And again, as I always try to do, here's the (rough, basic, unfiltered) HOWTO on what I did to get Kopia with garage up and running in my homelab. (don't worry, the keys and secrets have already been rotated)
https://codeberg.org/jwildeboer/gists/src/branch/main/2025/20251227-Kopia-Garage-Backup-Mac.md
Anyone using S3 (compatible) storage for incremental backups of Linux and/or MacOS machines at home? If yes, what backup software/tool? Open source preferred. My first unfocused research suggests to look at restic [1] and maybe just some scripting around rclone [2]. Any hints/links welcome! (Yes, this is part of my project to use local S3 compatible storage based on garage [3])
[1] https://restic.net
[2] https://rclone.org
[3] https://garagehq.deuxfleurs.fr
Thank you all for sharing suggestions and experiences! Keep em coming 😊 I am still in the "playing around" stage, but I just finished my first backup using Kopia [1] and I am quite impressed. Initial backup from a (WiFi connected) MacBook Air to my local garage instance (the old NAS box, a HP Microserver gen10) I did around 14GB in 10 minutes.
[1] https://kopia.io
Anyone using S3 (compatible) storage for incremental backups of Linux and/or MacOS machines at home? If yes, what backup software/tool? Open source preferred. My first unfocused research suggests to look at restic [1] and maybe just some scripting around rclone [2]. Any hints/links welcome! (Yes, this is part of my project to use local S3 compatible storage based on garage [3])
[1] https://restic.net
[2] https://rclone.org
[3] https://garagehq.deuxfleurs.fr
Hm. Does anyone have a little shell script that does this:
- Find all picture files
- Check if it has EXIF data and if yes, rename the file to YYYYMMDD_HHMMSS using the EXIF creation date
- If NOT, rename the file to the same pattern, but use the filesystem last modified date
I have `for img in $(ls *.[Jj][Pp][Gg] 2> /dev/null); do exiv2 -r'%Y%m%d_%H%M%S_'"$(tmp=${img%%.*};echo ${tmp##*_})" rename "$img" ; done`whicj does the first part, but fails when no EXIF
Hm. Does anyone have a little shell script that does this:
- Find all picture files
- Check if it has EXIF data and if yes, rename the file to YYYYMMDD_HHMMSS using the EXIF creation date
- If NOT, rename the file to the same pattern, but use the filesystem last modified date
I have `for img in $(ls *.[Jj][Pp][Gg] 2> /dev/null); do exiv2 -r'%Y%m%d_%H%M%S_'"$(tmp=${img%%.*};echo ${tmp##*_})" rename "$img" ; done`whicj does the first part, but fails when no EXIF
I wish I'd realised you can set up streaming replication with PostgreSQL with a newer version of the server at the receiving end much earlier. I've always matched versions, assuming I had to, when I could have saved time on later upgrades by promoting the replica after a sync.
A recurring theme in my professional life is finding myself needing to solve one of those little puzzles where you slide square tiles around in a frame, but this one doesn't have an empty space in it.
Dear software companies: You can put (much) more information in error messages. You're not protecting your delicate users from extraneous information—they're just going to ignore everything on the screen anyway—all you are doing is making it hard for your users' support staff. #SysadminLife
Dear software companies: You can put (much) more information in error messages. You're not protecting your delicate users from extraneous information—they're just going to ignore everything on the screen anyway—all you are doing is making it hard for your users' support staff. #SysadminLife
Should probably have put my management VPN address range on the sshguard allow list, shouldn't I?
Hmm one of the drives failed in my server cluster at home and I didn't get a notification about it.
I guess that I need to spend some time this weekend working on server monitoring and alerting, which will make a refreshing change from work this week where I spent a lot of time working on-
I have added a Proxmox Backup Server installation to my cluster of computer out back and I have to say I'm pretty impressed how easy it's been to get a three-stage setup going with different retention rates, including shipping off a set to Backblaze B2 for long-term storage.
It's taking bloody ages to sync off-site though, I really need to get that link upgrade sorted before they drop my upstream to 20 meg urgh.
You survived another week of patching, provisioning, and panicking.
Breathe deep, you've earned some quiet logs.
You survived another week of patching, provisioning, and panicking.
Breathe deep, you've earned some quiet logs.
The outcome of today’s #SysadminLife event:
![Mike [SEC=OFFICIAL]](https://static.chinwag.net.au/accounts/avatars/000/000/001/original/b02e2e0cd4f0ffe5.jpg "Mike [SEC=OFFICIAL]")
