Why introduce complexity when torrents and DNS are right there? What if dns pointed to something that was never online? let’s visit nonexist.whiting.dev!
I have a backup SSD that I keep on my person (even when I don’t have my laptop). This is essentially my use-case for when the laptop dies.
Recently, I’ve moved my restic backups from locally connected to the laptop, to connected via SFTP across to a raspberry pi. This presents an interesting challenge:
backup script: add it to cron, something like once every */3 hours
#!/bin/bash source some_env echo "backup ssd on $(date)" | tee -a restic.log if [[ -f .lockbackups ]]; then echo "not running - locked by other cron" | tee -a restic.log exit fi # magic folder check: if ssh user@server '[ ! -d /mnt/ssd/path_to_backup ]'; then #if [[ ! -d $RESTIC_REPOSITORY ]]; then echo "SSD not present" | tee -a restic.log exit fi touch .lockbackups restic backup --exclude-if-present .nobackup --exclude-file /etc/restic-exclude / | tee -a restic.log rm .lockbackups /etc/systemd/system/mnt-ssd.mount where $USB is your drive’s name. UUID=$(sudo blkid $(sudo blkid --label $USB) -s UUID -o value)
Read more
Should keep code snippet explanations permanent on whiting.dev
AWS: “stop doing undifferentiated heavy lifting” - meanwhile, to setup a simple app in AWS… yeesh. Something I read earlier, cloud needs to address 3 deficiencies: complexity, compliance, & cost.