SSD suddenly failing but shows 100% health, or something else?

spixelspixel · 16 Mar 2018 at 13:08

My PC suddenly froze today, could only click certain things but no application icons in taskbar would open.

Upon shutting it off and rebooting I was met by an error 'unmountable boot volume' Restarted and it tried to repair but failed. After several on/offs it eventually booted into safe mode where I tried chk disk and sfc /scannow. Both failed but tried a few more times and eventually worked.

Now every time I booted into windows after the first loading screen, it would show another loading screen with just the spinning circle for around 10 seconds. When I go into event log its showing lots of reported bad sectors on the harddrive so I assume it keeps doing some kind of scan on boot. Now the pc is still freezing every now especially upon first boot, everything is acting laggy and strange and sometimes the pc just locks up.

Would you say its the ssd or what? I tried everything except repairing the mbr as I cant seem to get it to work from the usb installation.

I'm hoping its just co incidence but I installed a new graphics card yesterday.

Anyone have any suggestions to try

If it is the ssd, proof these kind of programms are completely unreliable

mrbazmondo · 16 Mar 2018 at 13:58

Is the SSD firmware up to date?
Before it went wrong was there any power outage, or was it hard powered off in any way rather than a clean shutdown?

Some older SSDs, most notably Intel ones, when faced with an immanent power outage didn't re-prioritise and kept on housekeeping (background trimming blocks and wear levelling) rather than writing outstanding data that needed to be written.
The net result is the drive goes to write something, reads something like 2 meg chunk, mixes in the new data, goes to write it to a new location and doesn't get around to it, or fails to update the metadata that states where the block now resides before power loss.
When reading the block affected you get zeros - blank.

I had this happen with a customer on an old Intel SSD. They lost the first bit of the disk which had the bootloader on it. Newer drive don't have this problem.
I have a Samsung 840 Pro - It used to magically disappear from the system every now and then. After updating the firmware its been rock solid for years.

I would copy off all data you care about, update the firmware and reinitialise (full trim) the whole drive, then reinstall.

APM · 16 Mar 2018 at 13:59

Pull the graphics card out and try it again.

spixelspixel · 16 Mar 2018 at 14:04

mrbazmondo said:
Is the SSD firmware up to date?
Before it went wrong was there any power outage, or was it hard powered off in any way rather than a clean shutdown?

Some older SSDs, most notably Intel ones, when faced with an immanent power outage didn't re-prioritise and kept on housekeeping (background trimming blocks and wear levelling) rather than writing outstanding data that needed to be written.
The net result is the drive goes to write something, reads something like 2 meg chunk, mixes in the new data, goes to write it to a new location and doesn't get around to it, or fails to update the metadata that states where the block now resides before power loss.
When reading the block affected you get zeros - blank.

I had this happen with a customer on an old Intel SSD. They lost the first bit of the disk which had the bootloader on it. Newer drive don't have this problem.
I have a Samsung 840 Pro - It used to magically disappear from the system every now and then. After updating the firmware its been rock solid for years.

I would copy off all data you care about, update the firmware and reinitialise (full trim) the whole drive, then reinstall.

I had to turn it off by holding down power button as windows wasn't responding properly, mouse was moving + could click on some icons but no programes would open.

spixelspixel · 16 Mar 2018 at 14:05

APM said:
Pull the graphics card out and try it again.

I tried it, does the same thing on boot, 2 loading screens + then in event log, lots of bad block errors. Maybe seems a bit more stable without the graphics card but mostly acting the same through the IGP.

APM · 16 Mar 2018 at 16:40

I had a SanDisk 480GB SSD that ran flawlessly for over 2 years.

I think I overutilised the capacity though because after I had used 80% of the space on it it began behaving in a similar fashion to how yours is now.

If I were you I'd get all my data off that drive asap,have you got a recent system image you can back up to a new drive from?

Looking at the stats in a previous screenshot I doubt you can RMA it?

Updating the firmware and re-formatting is always worth a try too.

APM · 16 Mar 2018 at 16:43

Also sounds like it could be a dirty bit,maybe worth a quick search.

Pocah · 17 Mar 2018 at 07:54

It sounds very much like the sort of problems you could get back in the old days with hdd's if you pulled the power plug on the PC. Rather it sounds more like corruption than an actual fault with the drive. I would re-frormat and re-install Windows. Take it from there..

Disco_P · 18 Mar 2018 at 01:17

First thing first is create a backup whilst you can.

Then I'd try a different sata cable and test.

Failing that a fresh install and test.

Failing both it's possible that the drive is dying.

I've never trusted SSDLife, it's so rare that a drive will slowly decline from 100 and die at 0%, more often than not drives will suddenly start behaving poorly and then die. Fairly sure that the last disk health check I performed on my Samsung SM951 showed 100% health with nothing noteworthy on the SMART data. Now my PC cannot post with it installed.

spixelspixel · 20 Mar 2018 at 21:11

Disco_P said:
First thing first is create a backup whilst you can.

Then I'd try a different sata cable and test.

Failing that a fresh install and test.

Failing both it's possible that the drive is dying.

I've never trusted SSDLife, it's so rare that a drive will slowly decline from 100 and die at 0%, more often than not drives will suddenly start behaving poorly and then die. Fairly sure that the last disk health check I performed on my Samsung SM951 showed 100% health with nothing noteworthy on the SMART data. Now my PC cannot post with it installed.

I ended up just getting a new ssd, crucial mx500 500gb

Sandisks own program was telling me the ssd was out of spare blocks and to replace (with a new sandisk ssd lol) but if I refreshed it about 50% of the time it would say the drive is okay, and other times would say out of spare blocks. Couldn't seem to decide. Its very strange how quickly the ssd went from working perfectly to all of a sudden having like 10,000 bad block reports in a few hours. There seems to be a fallacy out there that smart stats of the ssd are reliable and people actually think their ssd is perfectly fine for another couple of years based on what it says, reality is it could fail at any time. Those tests that show how many writes an ssd can last are bs as its not a real world test. Mine was probably damaged from excessive browser usage. Except for installing a few games and browsing I'd never write much to the ssd and always had ~100gb spare.

mrbazmondo · 20 Mar 2018 at 21:38

You might be able to revive it by under provisioning. Purge/trim/initialise the whole disk, whatever they call it in the sandisk tool, then make a partition leaving 20G or more unused. The blocks that would have been used by that unused section will then be available for it to use as spare blocks for the part of the disk you are using.
The partition you create - don't full format, just quick. You only want it to allocate blocks that actually have data on them.