Decide which data is actually important to you and what is merely a pain in the ass if lost.<p>An example of the former is family photos and videos.<p>If you don't make the distinction you'll inevitably slip up in trying to cover everything.<p>Where you go from here can vary, but a good rule of thumb is: "three is two, two is one, one is none" in terms of copies of data. Redundancy is your friend here.<p>Media taken on my phone is stored on a cloud account as well as the local device, and usually for backups I do the following in order:<p>- External hdd backup<p>- a Blu-ray/DVD backup that is stored at a different geographical location from my device.<p>- AWS Glacier in a region on the other side of the world. (You shouldn't plan to use this for disaster recovery, it's a last ditch hail mary restoration effort and priced as such. Super cheap to store, expensive to retrieve).<p>Again, swap out individual storage mediums/services with your own preferences, rsync, tarsnap, whatever. But the general outline remains.
Asked a few days ago, too: <a href="https://news.ycombinator.com/item?id=29863822" rel="nofollow">https://news.ycombinator.com/item?id=29863822</a>
I have a simple NAS setup at a family members house far away and backups are pushed to that nightly from an internal drive in my machine (internal RAID). I also just sitting on my desk have a drive that gets those same back ups.<p>And then I have a cloud service I pay for that only looks at certain drives / folders and backs those up as well.<p>One tool to automate all of this would be nice though - currently running several different ones
Daily snapshots of the system configuration, personal documents and databases sufficient to rebuild the system to where it was last night in less than an hour or so.<p>Daily backups of the 'write once and keep forever stuff' - videos, music, photos, etc.<p>Three copies of the above backups on external drives.
I have several large(ish) external USB HDDs, and every few days I cycle them around the machines.<p>* On each disk is a directory called "Backups";<p>* Within that is a directory for each machine on my home network;<p>* Within each of those is a directory for each time of backup;<p>* Those are named YYYYMMDDhhmmss;<p>* Every hour I run a script;<p>* If the most recent backup is on the same day, the script terminates;<p>* Otherwise the script runs an rsync;<p><pre><code> PREV=$( ls ${Machine} | grep ^20 | sort | tail -1 )
THIS=$NOW
rsync -avziPH \
--exclude ".cache" \
--link-dest=../$PREV \
/home \
${Machine}/$THIS
</code></pre>
Thus each disk has backups from each machine. Each backup is complete in and of itself, but for unchanged files has a hardlink to the previous version.