Diagnosing SSD failure

Forums › Laptops and mobile computers › Diagnosing SSD failure

mdf

1956 posts

Uber Geek
+1 received by user: 574

Trusted

Subscriber

Topic # 228890 28-Jan-2018 11:31

I have a Sony Vaio Pro. Was having frequent BSOD crashes and so did a full reset of Windows 10. BSOD crashes continued; I'd have a period where it would constantly crash every few minutes, but then other periods where all was well for days at a time.

Several BSOD crashes this morning (as I was trying to read instructions for diagnosing driver issues to stop BSOD crashes...) that culminated in "Your VAIO failed to start Windows. System recovery failed to start from the internal hard disk drive/solid-state drive." From there I could get into the BIOS (nothing obviously wrong there), but nothing else worked.

Some time later (as I was finding another computer to post this) I was able to boot into windows normally but immediately shut it down before breaking anything else.

Is this an SSD failure or potentially something else? Is there a way of better diagnosing this (I may be able to do something from Windows if it behaves)?

I'm happy to replace the SSD if that is the likely problem. Files etc. aren't an issue; everything important is backed up elsewhere.

According to the interwebz, Sony Vaio Pros use a PCIe SSD. I've never bought one of these before; is searching "PCIe SSD" sufficient or are there other things to take into account?

DS248

381 posts

Ultimate Geek
+1 received by user: 13

Subscriber

Reply # 1947817 28-Jan-2018 12:01

Can't answer your questions but I also have a Vaio Pro that had an SSD failure (at about 4 years). Advised no simple fix - has to go in for servicing. That was over a year ago. Decided to get new (non-Sony!) ultrabook instead. Will get around to fixing the Vaio "one day"

cadman

1014 posts

Uber Geek
+1 received by user: 572
Inactive user

Reply # 1947822 28-Jan-2018 12:11

I had similar issues with a 3 y.o. Lite-On SSD in a Dell for about 6 months. It was fine for weeks at a time and then periods of instability. I didn't realise at the time it the was the drive that was causing the problems. Fortunately I already had another drive I had installed and was gradually setting up and then one day the Lite-On completely died. I never liked Lite-On from the days of their notoriously flakey CD-ROM drives and this more recent failure served only to reinforce my bias.

timmmay

14138 posts

Uber Geek
+1 received by user: 2544

Trusted

Subscriber

Reply # 1947823 28-Jan-2018 12:14

How old is the computer? Under warranty?

Two things I'd try:

- Get memtest or similar, something that runs from a boot USB, and check the RAM. RAM is always the first thing I test. If it fails it could still be motherboard or similar.

- Put the SSD into another computer and run some kind of testing software. I don't have anything specific to recommend. What I did when I had disk problems (that turned out to be bad RAM) was use Teracopy (free) to copy massive files around, and make sure "validate" is turned on. That does a checksum of the source and destination files. I'd format the disk first, use Macrium Reflect or similar to back up beforehand if you need to.

AWS Certified Solution Architect Professional, Sysop Administrator Associate, and Developer Associate
TOGAF certified enterprise architect
Professional photographer

lxsw20

2184 posts

Uber Geek
+1 received by user: 661

Subscriber

Reply # 1947826 28-Jan-2018 12:25

Can you boot in to safe mode OK? Anything in the event logs?

It could well be SSD, but i wouldn't say it's narrowed down to that. Could be plenty of things. As said above, try memtest to rule out a RAM issue too.

Lias

3354 posts

Uber Geek
+1 received by user: 1835

Trusted

Lifetime subscriber

Reply # 1947834 28-Jan-2018 13:16

mdf:

According to the interwebz, Sony Vaio Pros use a PCIe SSD. I've never bought one of these before; is searching "PCIe SSD" sufficient or are there other things to take into account?

It appears that Sony have shipped both M2 Sata and M2 PCIE drives in those laptops, and even then PCIE ones come in 2 and 4 lane variants. So, yes more things to take into account.

Information wants to be free. The Net interprets censorship as damage and routes around it.

mdf

1956 posts

Uber Geek
+1 received by user: 574

Trusted

Subscriber

Reply # 1948059 29-Jan-2018 11:02

I ran memtest last night, and happily everything seemed to pass:

Result summary
Test Start Time 2018-01-28 20:22:30
Elapsed Time 2:58:48
Memory Range Tested 0x0 - 21FE00000 (8702MB)
CPU Selection Mode Parallel (All CPUs)
ECC Polling Enabled
# Tests Passed 48/48 (100%)

mdf

1956 posts

Uber Geek
+1 received by user: 574

Trusted

Subscriber

Reply # 1948060 29-Jan-2018 11:09

I've been looking through the event logs this morning, and there seem to be a tonne of errors and warnings (though it's just cr*pped itself again!)

I'm really not sure what I'm looking for.

Most likely (?) is a Critical level event, "Kernel-Power" - "the system has rebooted without cleanly shutting down first...". However, the next issue is "volmgr", "dump file creation failed due to error during dump creation". Not sure how to investigate this further.

There seem to be a lot of "DistributedCOM" - "the application-specific permission settings do not grant Local Activiation permission for the COM Server application with CLSID [digits] and APPID [more digits] to the user NT AUTHORITY\LOCAL SERVICE SID (s-1-5-19) from the address LocalHost (using LRPC) running in the application contained Unavailable SID...

There are also a lot of TPM errors, which I assume relate to this: https://sony-paa-pa-en-web--paa.custhelp.com/app/answers/detail/a_id/174243 . That timing would potentially be consistent with BSOD onset?

fearandloathing

87 posts

Master Geek
+1 received by user: 22

Subscriber

Reply # 1948072 29-Jan-2018 11:21

I’ve had great success in recovering spindle drives with SpinRite. There are some caveats in using it with ssd.

timmmay

14138 posts

Uber Geek
+1 received by user: 2544

Trusted

Subscriber

Reply # 1948088 29-Jan-2018 12:23

I had to use HCI memtest to find my bad RAM, memtest couldn't find it. HCI had to run for about 16 hours to find the problem. HCI runs in memory and tests 2GB RAM per instance, so you have to run multiple instances.

Good that memtest didn't find anything though.

AWS Certified Solution Architect Professional, Sysop Administrator Associate, and Developer Associate
TOGAF certified enterprise architect
Professional photographer

mdf

1956 posts

Uber Geek
+1 received by user: 574

Trusted

Subscriber

Reply # 1948182 29-Jan-2018 15:09

Hmm, further update.

I installed a whole lot of Sony "stuff", that included a hard drive diagnostic tool. Passed that fine. I've then also managed to boot into safe mode and run it for an hour or so under 100% CPU load using a windows 10 CPU and ram checker. No crashes.

I guess I'm therefore looking at some kind of driver incompatibility issue, rather than hardware failure?

mdf

1956 posts

Uber Geek
+1 received by user: 574

Trusted

Subscriber

Reply # 1949088 31-Jan-2018 10:21

Update: I ran the PC in safe mode for a while and everything seemed fine. Decided it was a driver issue and first port of call would be an absolutely clean install using recovery media. The recovery media let me format the SDD, but when it came time to install it decided to tell me that I didn't have a SSD/HDD installed. Oh oh.

I pulled the notebook apart (not entirely without incident), gave it a clean (though it was pretty dust free), pulled out and reseated the SSD. Then reinstalled windows from recovery media. It's been running fine for 24 hours so touch wood that was the problem.

Though not sure why a potentially loose SSD connection would be fine in safe mode?