3 ms·

I have just one of the early SSDs still functional, the original Intel 80GB SSD, its too small for anything now so its retired with about 60% of its life remain

I have just one of the early SSDs still functional, the original Intel 80GB SSD, its too small for anything now so its retired with about 60% of its life remaining. Everything else of that era of early SSDs died. The oldest SSD I have still in use is an Crucial M4-CT512 (https://www.crucial.com/support/ssd-support/m4-25-inch-support https://www.crucial.com/support/ssd-support/m4-25-inch-suppo...) and that was likely bought in 2013. It still works OK but its always had an issue with some writes taking a very long time. The younger ones like the Samsung 970 Pros still work as new as do the cheap chinese PCI-E 4.0 drives. I think the early SSDs weren't reliable at all but most of the recent stuff seems to be a lot better than the HDDs in the NAS which have a failure or so every 3 years.

alias_neo 2y ago
Anecdotally, I have a similar experience in my NAS/Server. Unlikely Backblaze, I don't write much to my drives, they mostly get some writes of relatively small backups once or twice a day and that's it. I keep all of my VM boot drives on NVMe SSDs. I've only had one, very early SSD (SATA, OCZ branded) fail on me since the beginning, and I run significantly more SSDs than I do HDDs, in my desktop, laptops, servers etc. In the NAS/Server (It's a high-spec Proxmox server doubling as a NAS) I have had a bunch of HDD failures, WDReds (CMR) and Seagate IronWolf drives, typically around 4-5 year mark. For use cases like mine, where they spend most of their time idling, I think SSDs are a clear winner, HDDs are wearing out just by idling. My next failure I'll be replacing the mirror that fails with SSDs.