Click on the ‘Detailed Information’ button and see what it says. There may be errors logged. Look for anything that doesn’t appear normal, such as reallocated sector counts (should be 0 or very close to it). Something has triggered the warning, it should be evident looking through the detailed info. I’ve never had an SSD fail so I’m not sure what kinds of errors these things throw when they give indication of impending failure.
One post mentioned updating drive firmware, so I started down that process:
[cbrake@ariel ~]$ sudo fwupdmgr get-devices
[sudo] password for cbrake:
WARNING: UEFI capsule updates not available or enabled in firmware setup
See https://github.com/fwupd/fwupd/wiki/PluginFlag:capsules-unsupported for more information.
Gigabyte Technology Co., Ltd. To be filled by O.E.M.
│
├─SSD 870 EVO 500GB:
│ Device ID: 602b0a6cc821d155208724f0e22f8d111542b74c
│ Summary: ATA drive
│ Current version: SVT01B6Q
│ Vendor: Samsung (ATA:0x144D, OUI:002538)
│ Serial Number: S6PXNZ0RB29592Z
│ GUIDs: 97fe3b7f-7d9c-553d-b3bc-365d46ba8a0b ← IDE\Samsung_SSD_870_EVO_500GB_______________SVT01B6Q
│ c8b943ea-8532-519a-afa8-55d93323a07d ← IDE\0Samsung_SSD_870_EVO_500GB_______________
│ d86b4839-1599-56b3-99b5-203ea834e2c2 ← Samsung SSD 870 EVO 500GB
│ Device Flags: • Internal device
│ • Updatable
│ • System requires external power source
│ • Needs a reboot after installation
│ • Device is usable for the duration of the update
But then it gave me:
[cbrake@ariel ~]$ sudo fwupdmgr update
WARNING: UEFI capsule updates not available or enabled in firmware setup
See https://github.com/fwupd/fwupd/wiki/PluginFlag:capsules-unsupported for more information.
Devices with no available firmware updates:
• SSD 870 EVO 500GB
• ST2000VX000-1CU164
• ST3000NC000
• ST32000645NS
• ST4000DM006-2G5107
No updatable devices
Anyone have thoughts on what brand SSDs to get? I’m glad to pay a little more for something that is reliable (thought I was doing that with Samsung …). I purchased this drive in Jan, 2022, so not even a year old.
I have had Toshiba, Samsung EVO and Western digital, So far they all are well. Toshiba runs rootfs on my builder and is 5+ years old now. Samsung Evo 850 (1TB) is the scratch and has been so for 5+ years and its on SATA interface so a bit slower but works reliably, WD Black (1 TB) is the latest one (3 years old) and its the fastest one since its NVME. I think Samsung EVO or WD are good value for money. btw. I use f2fs for filesytem
@cbrake you should be able to warranty that drive back to Samsung and get a replacement with so few hours. Those EVO 870 disks should have a few years of warranty on them.
If you see READ FPDMA QUEUED type errors, sometimes these can be poor signal quality on the SATA lines (I’m assuming this is a SATA drive). If you still see these kinds of errors when using a different drive you may try buying a new SATA cable.
Given some very recent disappointment with Samsung 800-series SATA SSD drives, I’m very hesitant to recommend them. But I have found Micron 5200 drives to be quite good. The 5200 series has been replaced by the 5300 series, which I presume are also quite good even though I’ve not had any personal experience with them. I’ve also had a very good experience with Intel S3710 drives, although Solidigm doesn’t seem that interested in SATA SSDs anymore.
I’m a convert to using datacenter grade SSDs now. The performance is simply good and solid. Consumer grade stuff costs less, but you definitely get what you pay for, especially if you care about write endurance or reliable write performance of small files.
The SMART warnings were timely, but not much warning – the disk just died completely and does not even detect on boot. Impressive that they gave me some warning before it died, but not much.
@cbrake hopefully you have backups or haven’t lost too much important data!
Yesterday I learned that the Micron 5400 series drives are actually quite competitively priced. If you only need 500GB or 1TB they don’t command that much of a premium over a consumer grade SATA SSD. Sometimes price is what matters, but if you can swing the money I’d definitely consider a datacenter-grade drive going forwards.
yeah, have backups for static files. Most project files are in Git, mail is on Imap server, etc.
Where do you buy Micron 5400s? It seems Amazon is geared more toward consumer grade stuff. BTW, has Amazon completely taken over the computer parts market? Places like Newegg, which I used to use in the past, seem to have completely lost it and have become. BH Photo seems like a good source for some computer gear, but they again are geared toward consumer.
This is interesting – installed new drive, and attempted to install Arch – first time, got HD I/O errors pretty quickly. Tried re-seating the SATA cable at the MB, and it went much further, but eventually failed. So, may have a cable/MB problem here – drive may still be good … to be continued …
Hooked the dead SSD up to other workstation – can’t get much to happen. Initially it shows up in lsblk, but as soon as I do something like fdisk -l /dev/sdb, it shows size of 0 and I get tons of kernel errors. Perhaps the motherboard caused enough errors that the drive shut itself down or something. Could also be a power supply issue where the supply to the drive is not stable. My best guess is the MB caused the failure though. If I could run a factory reset on the drive or something, may be able to recover it, but a quick search is not showing anything. Hopefully I did not roast the new drive …
Unless there was some kind of power supply problem on the motherboard I doubt the drive got roasted, probably just generally corrupted that hopefully it will still be usable after a repartition and creating new filesystems. Doesn’t help with recovery of the data on board unfortunately, but that might be toasted anyway. I have had USB memory sticks destroyed by desktops, but almost assuredly that was due to ESD when inserting the sticks or other power anomalies related to hot plugging. I think it might be possible for rogue software to wreck an SSD but I doubt this is what happened to you.
I’ve had this kind of failure happen with a USB connected flash drive. Sometimes I’m able to write data to it but always within a short time it goes bonkers and there’s tons of kernel errors reported for it. Wiping it, blkdiscarding it, repartitioning, nothing helps, it’s dead. You may have a similar kind of failure. Sadly there’s very little info publicly available to understand how/why managed flash devices die.