Conversation
Notices
-
@crunklord420 Hey, i am backing up personal files to mdiscs. I want somehow to be able to verify that the mdisc was burned successfully. I was thinking of creating a checksum of all the files and then comparing after burn to see if anything went wrong. Do you have a better suggestion? Or maybe a suggestion for a good way/program to generate the checksum?
-
@Saxophone3784 just normal sha256sum. Additionally you may be interested in creating parity files (PAR2) and burning it to the same disc in case the disc in damaged in a way where there's only partial corruption. PAR2 files function as checksums as well.
-
@Saxophone3784 I've never used mdiscs.
-
@crunklord420 Thanks, patity files sound good. Actually, another question, have you had experience with 100gb vs 25gb mdiscs? Whatevwr I found on the internet say that people often have problem with 100gb mdiscs
-
@Saxophone3784 I don't use optical media. I have a backup setup based around borg/borgmatic. My critical data is backed up like 6x times across multiple drives, onsite semi-air-gapped HDD, 2x offsite server, 2x on-person flash.
-
@crunklord420 Is there a reason if I may ask? do you just not use optical media or you prefer other brand?
-
@crunklord420 As I understand you just don't have a use case for optical backups. Is that correct? Or you see some issue w them?
-
@Saxophone3784 I had previously mentioned mdiscs because the idea is cool. It fills a special niche where you wanna like actually bury something in a cave somewhere.
In reality it's too expensive and cumbersome to use for day-to-day backups. I think about fires, earthquakes, floods. Total regional destruction. I want to be able to be naked on the other side of the planet and be able to get my files. Mdisc doesn't fill this need.
My protection against bitrot is through duplication and diversity. I was happy to pickup some used enterprise HDDs on deep discount. I don't need to worry if they're gonna fail because they're just yet another backup. Every time I get a new storage location it just gets added to my borgmatic script.
-
@Saxophone3784 also, ultimately, the answer is tape but you need true scale to justify tape. You can buy a really old machine for not too much, but the failure rate on them is probably very high. It's just too expensive. Datahoarders used to talk about shucking but I don't see a discount on external vs internal these days. Used enterprise seems like the best deal.