Precisely why I had the SSD (used for in-car navigation). Though, I have read of SSDs suddenly failing, more due to bad components than writing too may times hence the caution.

~2012. Sandforce does ring a bell.

Generally, yes, they are far more reliable. I have had a few SSDs fail within the first months of usage, but so far none due to old age/abuse yet. BUT, the big difference between SSDs and spinning rust HDDs, is that the SSD failures I've seen gave no warning. One day it was working fine. The next, it wasn't responding.

Always, no matter the technology -- Backup, backup, backup! How much is your data worth to you? How much $ would it cost you to get back to normal again? How much would you be willing to pay to get that data back? Usually the answer to that is far more than the cost of a small portable hard drive. I don't have an issue with using HDDs as backup drives, in fact that's what I use for offline backups. For most people, the "3-2-1" style of backup works well. ie: 3 total copies of your data... 2 on-site (one on the machine, one external drive), 1 off-site in case something happens.

What you're proposing has no possibility of working--it's impossible to run Windows from an external drive. You clone to an internal drive. (It's not hard but I would strongly suggest getting help from someone who is comfortable working on the innards.)

As for the difference between cloning and copying--a clone is simply a perfect copy. It must be done with a tool meant for the purpose (Macrium Reflect's free version will do it), it's not something you can do with Windows itself. You need to not only copy the files, but the partitions.

Happens to both SSDs and HDDs; almost all computer parts follow a bathtub curve to at least some degree.

For desktop drives where impact is rarely an issue, the controller board going can be just as sudden. For both SSDs and HDDs, poorly cooled systems are IME biggest culprit but some drives of either type will just die even with great cooling and no impact/shock. Mechanical failures not due to external forces aren't as frequently catastrophic on HDDs as they used to be -- some folks here will probably be old enough to remember when you had to park HDDs before shutdown :)

I'm very happy to be back in pure software development where hardware can be treated as a virtual resource in most cases, but I spent one of the most interesting parts of the HDD-SSD transition for data centers (2010-2015) in a couple of positions where I did a mix of development and IT management. I got far too familiar with the constant drip-drip-drip of HDD replacement rates.

IME, that's a fairly common failure mode for HDDs as well, both the "motor no longer spins up" mechanical failure and the "controller board dead, computer doesn't see drive."

Excellent advice. I've moved to SSD for most of my non-bulk backup (the bulk being ;the tens of terabytes of media on my server at home), but it's still a bunch more expensive to do so. One thing that isn't clear for either SSD or (modern) HDDs is the long-term viability of cold-storage data. The oldest still-running HDD I have was 28 years old last summer when I dug it up at my mom's house -- and still has intact from when I was in high school (although it took some work to did out of a dead laptop.) The oldest SSD I've got is 11 years old, and still has readable data on it. OTOH, bit rot is real, and modern SSDs use a much smaller feature size (and this a much smaller amount of captured electric charge to read) than the IIRC 50nm of that first-generation Intel drive, and the density of magnetic media has gone up very rapidly as well.

Optical media specifically designed for long-term data retention, or tape, are probably the most secure ways of very-long-term cold storage. These days of course, it's also just practical to copy everything onto a bigger drive every few years, and then if one old copy dies you have a newer one (although at a certain point, you need to worry about bit errors on copy.)

Not sure about MS's program, but for Macrium/Acronis/Clonezilla/Gparted, there's nothing wrong with cloning to an external drive and then swapping that drive (minus the external case) into the machine, or cloning off of that onto a different internal drive.

Also, with some poking, you can run Windows from USB. It's miserably slow (even on USB 3+) so it's basically only used for creating Windows-specific rescue environments, but it is possible.

My preferred method for data safety is a RAID server. With RAID 1 or 01 (10?) the important property is data redundancy. If a hard drive fails, it lets you know. The data is still safe on the other drive and you have time to replace the failed unit.

Some computers come with a RAID option built-in to the bios, but I prefer a hardware solution (RAID card).

I'm using an older dual-core Dell machine, and the connectivity is via an air-gapped GB network. The only drawback is that I have to be wired in (not available via Wi-Fi) which I consider a security feature. E.G., I don't complain that I have to use a key to open a safe :).

Anyhow, I've been doing it for years and have never lost a byte. YMMV

Like the old saying goes, there are two kinds of people: those who have lost data and those who will.

As any data storage person will tell you, RAID is not backup. If you get a power surge or controller failure or the second drive simply dies before you can replace the first one and recreate the mirror, you're out of luck. I use RAID 1 on my basement server, but I also back up to Crashplan.

Anyone know what's up with not being able to force-kill a process anymore? Task manager simply shows it as not responding. It's running two processes, Taskkill supposedly nailed one (it's still there, though), it says the other doesn't exist.

NEVER rely on Raid as a backup. I've said it on here before, but I've had 2 occasions where Raid has let me down.

1. One drive failed in a Raid 5 configuration. I replaced the drive, and the rebuild of the array started. The rebuild failed with a weird error and the whole array was corrupted. This was with a Raid card.

2. One drive failed in a Raid 5 configuration on a Synology NAS. I replaced the drive, and the rebuild of the array started. A second drive failed during the rebuild, so everything was lost. When I was rebuilding from scratch, the 3rd drive failed! At this point I realised there was something amiss - so I replaced all the drives and the NAS they were present in.

Of course, I have everything backed up to multiple sources, so it wasn't an issue other than the time and expense of replacing drives and the NAS :D

I do it via 'End Process Tree' from Task Manager. I've not noticed any issues.

really?

-David

I've made some good money recovering data from failed RAIDs, including some larger organizations which should have known better.

Always, always, always, buy your drives from different batches, especially if they're going in the same system. A client of mine had bought a brand new Dell server, fully loaded with 8 drives from Dell. Within the first 2 months they experienced multiple drive failure, with the second drive failing during the rebuild. Buying drives from different batches is actually quite simple -- buy them from multiple vendors and over a period of time if you have the luxury of doing so.

Also, make sure whatever devices you use, you have spares of. If it's tape, make sure you have an extra tape drive. If you do use hardware RAID, get ahold of a spare RAID controller.

I've been doing software-only RAIDs for many years. There used to be a huge performance hit for it, but no longer. Being able to used mismatched drives and any hardware? Priceless. You can buy used enterprise-grade hardware for almost peanuts off eBay. Used 3TB HGST drives are $25 each, or 12TB for $100. Pair it up with FreeNAS and you've got something that's near-bulletproof.

Supposedly you can mess with the system at a very deep level to make it work, but with out-of-the-box Windows the USB drivers aren't loaded soon enough, it tries to read the system drive before it has the ability to see it.

These days that might be tolerable. With software RAID you'll have a rebuild every time Windows pulls a BSOD. Such crashes used to be reasonably common.

Why would you need to do that? You can move a Windows Storage Space between systems without any need to rebuild.

there's something called wintousb to do the install to an external drive.

-David