For the second time, I have removed a disk from its purchased enclosure (the first was IEEE1394/FireWire, yesterday's was USB 2). The result both times is that the disk was exhibiting wonky behavior in the enclosure, either outright throwing errors or hearing the read/write heads being repositioned a lot of times (you experienced computing folks know exactly what I heard), and after being extracted, work just fine on its own. The lamentable part is both of them were intended as backup disks, and the next-to-last thing you want to go wonky is your backups (with your first thing you don't want to go wonky is the computer itself). A case-in-point follows.
Almost ever since I got my 2 TB Seagate FreeAgent, it would have that tell-tale "I'm having trouble reading the platters" thonking of the heads. However, Seagate will not take warranty claims unless their utility (ugh...Windows only, .NET 4 requiring) tests it and the utility pronounces it defective (or put another way, it sounded like if you tried sending it in for a warranty claim without their utility finding it defective, you'd be economically responsible, not Seagate). At the time, I should have taken that as a hint that lots of folks found these disks dodgy, probably with the same almost unmistakable "help! I'm having read problems!" head thrashing, and had their warranty claims shot down. After all, why would Seagate even have to caution people about that on their site? But I digress. Recently I pressed it into service as the storage for MythTV recordings attached to a Raspberry Pi. Finally, this past week I had had enough of sitting there during recordings, and at times during viewings, of hearing the thwacka...thwacka...thwacka of repositionings/retries. I bought a 2 TB Toshiba USB disk (essentially it's a laptop drive with a case and a USB to SATA converter). While copying the MythTV videos from one disk to the other, there were plenty of times I heard that head banging. I then put the Toshiba drive in Myth service, with seemingly the only detriment being that if you plug it into the running Pi, it makes the voltage dip below threshold (making the red LED go dark for half a second or so) during the time the disk motor spins up. (After all, the Pi is only USB 2, the Toshiba is a USB 3.x unit, therefore has those higher allowable current draws.) I then proceeded to tear apart the Seagate case.
After putting the Seagate drive on a Sabrent USB 3 disk converter (turns out it's a Barracuda LP at heart), I did the same rsync copy where I heard the clackity-clacks before, but this time, there was no such noise. Soooo.....did it remap sectors and now it was getting a clean read? Was it overheating in the enclosure? Was the Seagate USB converter board wonky? Was the power supply unstable? Without some professional diagnostic tools, and maybe a clean room, it will be VERY difficult to tell for sure, but I'm guessing it's unstable power to the drive, like the 160 GB LaCie FireWire drive that preceeded it.
That's OK...the FreeAgent power suppy is being used for a different purpose, on some audio gear, where it doesn't seem to make a difference. Whatever glitches it might have had do not seem to be audible. My money's on the traces on the USB to SATA converter board just weren't up to scratch, and didn't provide enough stable current for the disk.
English is a difficult enough language to interpret correctly when its rules are followed, let alone when the speaker or writer chooses not to follow those rules.
"Jeopardy!" replies and randomcaps really suck!