[HN Gopher] Thoughts on "The Theory and Craft of Digital Preserv...
___________________________________________________________________
Thoughts on "The Theory and Craft of Digital Preservation"
Author : todsacerdoti
Score : 39 points
Date : 2021-07-19 07:11 UTC (1 days ago)
(HTM) web link (ajroach42.com)
(TXT) w3m dump (ajroach42.com)
| mftb wrote:
| Long article, so I had to bookmark it for later, but I'm glad
| this made it to the front page of HN. Digital preservation is
| important to me.
| rambambram wrote:
| Also bookmarked, but also read 80% of it. The author is talking
| terabytes of data, but I see some overlap with my own humble
| 'backup flow' (as I call it).
|
| In my case, I'm doing a couple of gigabytes max. I made a
| distinction between my essential files on one big USB stick
| (what I need to get going again in case of a broken computer),
| and the bigger files like audio and pictures on another big USB
| stick. I also use rsync (mostly) and a mysqldump per database,
| all nicely in a bash script that I can quickly run but also
| quickly change. The USB stick is FAT formatted and the contents
| are not zipped or otherwise compressed. Just plain files as I
| saw them on my origin computer, easily accessible with other
| computers than my Ubuntu machines. And the most important part,
| I can quickly verify the success of the backup with my own
| eyes. Is that important file also backed up? Yes, it is!
|
| I also decided on and implemented a 'backup flow' wherein I: -
| set up new computers with the least amount of files copied from
| the previous machine (this is done manually, and usually I take
| one week in which I gradually transfer the files I need); -
| just keep all files from the old computers on my backup USB
| stick, so I have access when I need to and can forget it for
| the rest; - just backup the new computer to it's own 'root
| folder' on the USB stick, so old and new computers can have
| their backups on the same USB stick (gives me more oversight
| and makes it more portable for me).
|
| Just my two cents. I'm very curious how other HN'ers handle
| their backups!?
| state_less wrote:
| I'm starting to wonder if I've had a hobby of collecting
| information and knowledge this whole time, and just now
| realized what I've been up to.
|
| I'm sorry if this sounds like it's coming out of left field,
| but I really love you all, for all that you've shared with me.
| Thank you.
| adultSwim wrote:
| I think preservation should be verifiable. ZFS, AWS S3, etc are
| great and should be used, but not relied upon to prevent data
| corruption. That should be a separate step on top of the storage
| layer.
|
| We can also do better than hashing by using error correcting
| codes. If a file doesn't match a stored hash, you have to check
| other backups to see if the file went bad or if the hash went
| bad. ECC can do this without checking backups and has the ability
| to "heal" a damaged file (see Reed-Solomon, which is used in RAID
| 5/6).
|
| Parchive/par2 is an off the shelf replacement/supplement for
| checksums.
___________________________________________________________________
(page generated 2021-07-20 23:03 UTC)