Always, always backup. And frequently! Don’t trust your local harddrive (especially if it’s a device you frequently take with you), don’t trust flashdrives, don’t even trust your local fileserver if it doesn’t have built-in backups (and even if it does, check that those backups actually work). If it’s not saved on at least two physical places (two drives in the same PC/server count, but it’s sketchy on its own), it’s not backed up!
I just scatter mine under the fingernails of multiple unhoused individuals throughout the city. It’s a bit of a pain, but it’s peace of mind. I’m thinking of expanding into microfiche hidden in fortune cookies next.
One backup copy isn’t enough anyway! The more the merrier, just make sure that enough of it is automated that your backups don’t get stale, and ideally stagger the timings so you don’t immediately overwrite all the automated backups with trash data once something goes wrong.
At one point I accidentally deleted a file, but I could conveniently copy it from the copy in my fileserver that automatically gets updated every two weeks.
I only do manual backups, but I’m the kind of person who does it multiple times a day anyway - whenever I do a major edit on a work or hobby file - just for my peace of mind :)
And yes, only “airlocked” backups. I manually use FreeFileSync to mirror my files to a local backup folder on another HDD (I have multiple paired folders set up inside that, so FFS doesn’t have to check tens of thousands of other files if I only edited a particular project that day), and keep only that synced to Google Drive. So if either the active local copy or Google Drive is corrupted or lost, the file is not automatically lost on the other end. I also found it a neat surprise that Google Drive retains past versions of files, it came handy a few times.
Lucky me, I recently acquired two 4TB hard drives, and I’m making a point to use one as my active backup, and every few months or so, clone that drive to the other one and swap them, just in case one ever fails ya know…
Doesn’t swapping increase the chance of total failure?
You’re basically using them equally, which makes it more likely that the surviving hard drive failes while copy the data to the future brand new replacement drive.
(This is obviously assuming, that storing a drive is different to you using the drive and that both drives will fail around the same time)
Meh, anything can happen at any time, I’ve come to accept that.
These are basically brand new surveillance grade hard drives, which means they’re designed to run 24/7 and run cooler than consumer grade drives. Neither one has any bad sectors, and they both purr quieter than kittens.
I always keep one completely offline and totally disconnected, sitting on a shelf, save for the occasional clone day where I have both connected.
I figure the chances of both failing at once are on an astronomical scale.
RAID isn’t a backup. It only protects against one mode of data loss, disk failure, which is probably the one the average user should be least worried about.
My CPU crapped out. And the wire leading to the thing that goes ‘beep’ when you turn on cpu was broken.
In trying to diagnose the CPU issue, I had to turn computer on and off a lot.
Somehow, doing that on and off repeatedly corrupted the hard drives. So raid doesn’t protect against problems like that or power spikes that fry things.
Its a little cylinder with 2 wires leading to the mobo. Not for error codes but for the ‘beep’ that happe is when you turn on computer. Is that not a thing any more?
Nope. At least, not that I’ve seen even slightly recently. I got into PCs ~15 years ago, and they were already becoming a lot less common then. It probably still exists in some niche way, that’s usually how it goes. Maybe HP still uses them or something like that.
Sorry if any of this is stuff you already know: The beep is a POST code- power on self test. That beep when you turn on the computer is basically the computer saying, “everything started correctly, from here on it’s probably a software problem.”
If there is a problem and your motherboard can figure out what it is- bad cpu, bad ram, no video, etc- it gives a POST code via the little speaker. It’s a nice troubleshooting tool, because a lot of the time the hardest part of the fix is figuring out what part is the problem.
Redundant network storage is cheap and available. If you’re a little tech savvy, one of those and a cheap hosting plan accomplishes two copies local, and one remote.
Always, always backup. And frequently! Don’t trust your local harddrive (especially if it’s a device you frequently take with you), don’t trust flashdrives, don’t even trust your local fileserver if it doesn’t have built-in backups (and even if it does, check that those backups actually work). If it’s not saved on at least two physical places (two drives in the same PC/server count, but it’s sketchy on its own), it’s not backed up!
3-2-1 Backup: 3 copies, on 2 types of media, 1 of which is offsite.
I just scatter mine under the fingernails of multiple unhoused individuals throughout the city. It’s a bit of a pain, but it’s peace of mind. I’m thinking of expanding into microfiche hidden in fortune cookies next.
I convert all my files to wavforms and teach starlings the songs
Hard to take the photo of it, but I backed up your comment to a CD:
Software used: https://github.com/arduinocelentano/cdimage
“Well, I don’t know what I expected“
I do the same thing but inscribe them onto clay pottery.
Don’t forget to draw your backups in the sand…
Uh, you should maybe read up on the latest trump executive order. Those unhoused fingernail storage units are likely to all end up in the same place.
If your house/office burns down, all your data is lost. At least one backup should be off-site!
One backup copy isn’t enough anyway! The more the merrier, just make sure that enough of it is automated that your backups don’t get stale, and ideally stagger the timings so you don’t immediately overwrite all the automated backups with trash data once something goes wrong.
At one point I accidentally deleted a file, but I could conveniently copy it from the copy in my fileserver that automatically gets updated every two weeks.
I only do manual backups, but I’m the kind of person who does it multiple times a day anyway - whenever I do a major edit on a work or hobby file - just for my peace of mind :)
And yes, only “airlocked” backups. I manually use FreeFileSync to mirror my files to a local backup folder on another HDD (I have multiple paired folders set up inside that, so FFS doesn’t have to check tens of thousands of other files if I only edited a particular project that day), and keep only that synced to Google Drive. So if either the active local copy or Google Drive is corrupted or lost, the file is not automatically lost on the other end. I also found it a neat surprise that Google Drive retains past versions of files, it came handy a few times.
Lucky me, I recently acquired two 4TB hard drives, and I’m making a point to use one as my active backup, and every few months or so, clone that drive to the other one and swap them, just in case one ever fails ya know…
Doesn’t swapping increase the chance of total failure?
You’re basically using them equally, which makes it more likely that the surviving hard drive failes while copy the data to the future brand new replacement drive.
(This is obviously assuming, that storing a drive is different to you using the drive and that both drives will fail around the same time)
Meh, anything can happen at any time, I’ve come to accept that.
These are basically brand new surveillance grade hard drives, which means they’re designed to run 24/7 and run cooler than consumer grade drives. Neither one has any bad sectors, and they both purr quieter than kittens.
I always keep one completely offline and totally disconnected, sitting on a shelf, save for the occasional clone day where I have both connected.
I figure the chances of both failing at once are on an astronomical scale.
Why not just use a RAID?
RAID isn’t a backup. It only protects against one mode of data loss, disk failure, which is probably the one the average user should be least worried about.
My CPU crapped out. And the wire leading to the thing that goes ‘beep’ when you turn on cpu was broken.
In trying to diagnose the CPU issue, I had to turn computer on and off a lot. Somehow, doing that on and off repeatedly corrupted the hard drives. So raid doesn’t protect against problems like that or power spikes that fry things.
I haven’t seen a PC that would actually have audible post codes in a very long time. Nowadays it’s usually LEDs, or a very simple little display.
Its a little cylinder with 2 wires leading to the mobo. Not for error codes but for the ‘beep’ that happe is when you turn on computer. Is that not a thing any more?
Nope. At least, not that I’ve seen even slightly recently. I got into PCs ~15 years ago, and they were already becoming a lot less common then. It probably still exists in some niche way, that’s usually how it goes. Maybe HP still uses them or something like that.
Sorry if any of this is stuff you already know: The beep is a POST code- power on self test. That beep when you turn on the computer is basically the computer saying, “everything started correctly, from here on it’s probably a software problem.”
If there is a problem and your motherboard can figure out what it is- bad cpu, bad ram, no video, etc- it gives a POST code via the little speaker. It’s a nice troubleshooting tool, because a lot of the time the hardest part of the fix is figuring out what part is the problem.
Redundant network storage is cheap and available. If you’re a little tech savvy, one of those and a cheap hosting plan accomplishes two copies local, and one remote.
I’m using a laptop with external USB adapters.
Check my comment history though, my very last comment to another post made a silly reference to RAID…