While copying a large amount of data (close to 4TB) from an external hard drive to a newer one I noticed that the drive that was being copied from got ejected in two separate instances at the same exact point in the transfer. Digging around a little it seems like a particular file refuses to copy over and gets stuck while copying. All other files seem ok as of now.
I was wondering if this points to bad sectors on the drive or to some other issue. Disk Utility seems to think that the drive is fine when I run First Aid on it. I’d be grateful to get tips on what I should do with the drive. It would be great if it can be salvaged. If you guys think that it is failing and can no longer be trusted then I will discard it after copying the data over.
FYI, the drive is a 4TB WD My Passport that is a little over three years old and is formatted as Mac OS Extended (Journaled). It has been used lightly, around once or twice a month at most. I am using a MacBook Pro running Mojave to access it.
No, when I try to copy the file in question manually it just stalls. When I set all the folders to copy overnight I woke up to “Disk not ejected properly” messages. Everything up to that file had copied over though so it looks like copying that particular file eventually crashes the drive.
Unfortunately it has happened again with another file :-/
EDIT: Eventually all but the two files that crashed the drive copied over. I am now checking the hard drive for bad sectors using sudo fsck_hfs -fS /dev/disk2s2
Personally no matter what you find out in terms of bad sectors I’d consider the drive to be suspect at best. I wouldn’t necessarily stop using it but I would use it only as part of a more complete backup system.
My personal goal is that any given working or backup drive shouldn’t cause me a major problem if it crashes. I’m willing to lose a days’ work but that’s about my limit.
I’ve read, in one place anyway, that even one bad sector can be an indication a drive is going to fail.
(At the risk of sounding like an advertisement) after I buy a new hard drive I “certify” them with SoftRAID, which writes random data to each part of a disk and reads it back, and finishes by zeroing the drive. 3 passes takes a day or two, but to my mind seems worth it.
Modern drives with high data density and spin rates have definite lifespans. A while back Backblaze looked at the drives it was using and found a jump in failure rates of between 9% and 17% between years three and four of their drives running 24/7, with an average annual failure rate of almost 12% by Year 4.
So, the sudo fsck_hfs -fS /dev/disk2s2 command has finally stopped running with this ominous message:
Scanning entire disk for bad blocks
Got a non I/O error reading disk at offset 2505421357056: Device not configured
It looks like the moment the scanning gets to a certain point it makes multiple tries, fails and the disk ejects itself, which is what happened during the copying phase as well. I am tempted to run fsck_hfs -fy -S /dev/disk2s2 command to see if it salvages the drive. Since only two files out of a few thousand are impacted I wonder if it makes sense to use it for non-critical tasks. What do you guys reckon?
Longevity doesn’t necessarily depend on wear and tear. I have two local backup drives I each use every other week, and I replace them within 5 years regardless of their condition.
The average life of a hard drive depends on a lot of things, like the brand, type, size, and interface method, but you’re looking at about four years on average. Online backup service Backblaze studied the drives in their infrastructure and found about 80% of them survived for four years. Of course, that also means 20% didn’t and failed sooner, most of those in the third year of use. Similarly, the brand of drive you use makes a difference.
I would be more enthusiastic about recommending it, but it’s pricey software. I got mine a long time ago and have kept it updated because of its usefulness to me, but OTOH I doubt someone could justify the price just to test disks.
If you’re really interested in reusing the drive, I’d swap cables and how it gets its power – but of course as the disk seemingly is consistently failing at a certain point in a copying process this doesn’t suggest the more usual USB hiccup or power drop/spike induced reason for the error.
So, the fsck_hfs -fy -S /dev/disk2s2 command didn’t go through either and the drive eventually disconnected. Interestingly, I was able to access one of the two files that I couldn’t earlier. I have a few questions that I’d love to get answers to:
Can I trust the files that I was able to copy off of the drive or could they be suspect?
What is possibly the issues with the drive? Are these bad sectors that I am running into because of normal wear and tear after three plus years of usage?
Is there anything I can do, in addition to changing the cables, to salvage the drive for non-important use?
Impossible to say. As @bowline implied, there’s really no rhyme or reason as to why a drive might fail.
You might try connecting the drive through a powered hub, although I agree with @timo that it’s kind of a long shot. I can’t think of anything else you could do.
