Leap 15.0
KDE desktop
Solid state drive using ext4
I have a folder on this drive that has 31 files in jpg format. When I try to open any of them, using any of several applications, I get an error saying Loading meta information failed.
Using Dolphin, and checking the Preview button, I can see thumbnails and the files seem to be in order.
Is there a way to recover these files?
Bart
What does:
file filename.jpg
show for one of the offending files?
bart@UNIVAC:~/MRH/Photos> file DSC_0935.JPG
DSC_0935.JPG: ERROR: cannot read `DSC_0935.JPG' (Input/output error)
That looks like a problem accessing the file. If this is on a physical disk, it suggests are hardware problem or a messed up disk format. If it is a network share, then it could be a network access problem.
Well, it’s worse than I thought. The files I mentioned are on a partition on a separate drive that is mounted in my home directory. It seems everything in both partitions on that drive is unavailable. Yet, the thumbnails show properly when I use Dolphin Preview in any of the folders.
The drive in question is a Samsung EVO 850 500G. I have another on this machine with two partitions and it is working correctly.
Using Disks, the drive seems good. I looked at SMART data and it shows no errors.
Bart
The thumbnails show correctly because they’re cached (normally on your home partition at ~/.cache/thumbnails) and were created, I assume, before the problem arose.
So, does that mean the list of directories and files are also cached? I open dolphin and can move around and see everything on the drive.
At the command line, I tried to copy… Oh wait… Here’s what I did.
bart@UNIVAC:~/Modelrailroad> cp Weathering_Track.pdf I-hope-this-works.pdf
cp: cannot create regular file 'I-hope-this-works.pdf': Read-only file system
bart@UNIVAC:~/Modelrailroad>
That read only part doesn’t seem right. I’m at a loss. Think I could dd to another drive?
Most of this stuff is backed up BUT I neglected to add one directory to my backup process and don’t you know, this one I absolutely need the files!
If there are serious errors, the system mounts it read-only. Yes, you could try copying, but with no guarantee of success.
It’s beginning to look like a corrupted file system and/or failing drive.
Anything shown in the journal? …
… or the output of:
sudo smartctl -a /dev/sdX
where “X” is the relevant device
Maybe boot from a live CD/USB and run fsck on the drive.
Response to smartctl is:
bart@UNIVAC:~> sudo smartctl -a /dev/sdd
smartctl 6.6 2017-11-05 r4594 [x86_64-linux-4.12.14-lp150.12.61-default] (SUSE RPM)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org
Short INQUIRY response, skip product id
A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options.
bart@UNIVAC:~> sudo smartctl -a -T permissive /dev/sdd
smartctl 6.6 2017-11-05 r4594 [x86_64-linux-4.12.14-lp150.12.61-default] (SUSE RPM)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org
Short INQUIRY response, skip product id
=== START OF READ SMART DATA SECTION ===
SMART Health Status: OK
Current Drive Temperature: 0 C
Drive Trip Temperature: 0 C
Read defect list: asked for grown list but didn't get it
Error Counter logging not supported
scsiModePageOffset: raw_curr too small, offset=119 resp_len=1 bd_len=115
Device does not support Self Test logging
bart@UNIVAC:~>
I guarantee the temperature is above 0 C
If I run smartctl on sdc, I get a ton of information. It does look like I lost this drive.
I’ll try to boot with a live disk and see if I can run fsck.
I’ll probably get only one chance at a copy. What parameters or arguments would you recommend?
Bart
Often smart is turned off on SSD drives since it adds to the write usage and thus shortens drive life. You can use smartctrl to turn it on (man smartctl) but there will be no history but you can then run the tests
Before doing that it might be worth powering down the PC and checking that the SATA cable to the drive is seated correctly, although in this instance I don’t really think that’s the cause of the problem.
Personally, I’d run fsck with the -V option to give verbose output.
Well, that didn’t work out well:
UNIVAC:~ # smartctl -s on /dev/sdd
smartctl 6.6 2017-11-05 r4594 [x86_64-linux-4.12.14-lp150.12.61-default] (SUSE RPM)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org
Standard Inquiry (36 bytes) failed [Input/output error]
Retrying with a 64 byte Standard Inquiry
Standard Inquiry (64 bytes) failed [Input/output error]
A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options.
Here’s an update:
The drive in question is sdd.
It is partitioned into sdd1 and sdd2
Both partitions have the same problem.
After shutting down and upluging the drive cable, I repowered up.
Using Dolphin, I did a copy of the the entire sdd2 partition to a new directory on another drive.
the entire directory structure was created but without any individual files.
Using Dolphin again, id went into each subdirectory and selected all files and copied them to the corresponding directory.
this resulted in some files being successfully copied!
The only Live disk I have is a copy of 13.1. But it’s not set up for UEFI. My computer refuses to boot without UEFI Can I unmount sdd and run fsck on that drive from my regular system?
That might work.
From what you have posted, it is possible that the disk surface is fine but the control circuitry is failing. It is hard to deal with disks where that happens. Maybe leave it unplugged for a while so that it can cool down. Then see what happens when you plug it in. Maybe set it to mount read-only, and see if there is anything that you can recover.
I have come to the conclusion that you have figured it out. I found that if I power off the computer, then reboot it, I can copy up to 6 files from the drive to another drive, one at a time, slowly, before it goes into error mode. This is going to be a long, slow process, but I think I can get what I really need off it. Then, it’s to the range to see if I can hit it at 500 yards. >:)
Thanks all for your input. I promise I’ll figure out a better backup system! It’s been over 12 years since I’ve had a drive failure and I guess I got a little careless. This drive is probably only 6 or 7 years old and I guess, being solid state I didn’t think it would wear out. Perhaps I should think about replacing all the drives in this machine.
By the way, unmounting it and running fsck -MV /dev/sdd2 worked just fine. After a reboot it shows the drive is good. After copying 6 files, it shows the drive as bad.
Bart
The only thing I would add, if you’re not doing it already, is to prioritise what you copy, take the most important files first…
Hope all goes well and you recover your data.
On Sat, 08 Jun 2019 05:36:04 +0000, montana suse user wrote:
> Then, it’s to the range to
> see if I can hit it at 500 yards. >:)
>
Way back when even 1GB drives were rare, I had one that I had to remove
from the chassis and chill before restarting. Last ditch was to move the
computer to the kitchen, stick the disk in the fridge with a long cable
and let it cool off then try to read it. Once it died doing it this way,
it was GONE! Wouldn’t even spin up but I got most of what I needed off
first.
I’d be surprised if most anytime regulars of OS/2 forums and/or mailing lists hadn’t used the freezer for a HD at least once. I certainly have. 
Along the same line, I’d try taking the plastic cover off the SSD if it has one, and place the drive such that I could have a fan blowing across an ice pack at it before trying more file copies. Oftentimes temperature makes a difference.