Unofficial archive. This is a read-only, community-run archive of the Apple Support Community. It is not affiliated with or endorsed by Apple Inc.

Can anyone give any advice on erasing (completely) a time machine back-up drive?

I have a back-up G-Drive that is full, and on the last back-up it was 'making space' by erasing the first back up.


Later when I transferred the backed up files to a new macbook, some were corrupted somehow (by the overwriting process of the back-up drive?), so the set up on my new macbook was weeks of resolving many corrupt files and half the things it couldn't do properly.


I've now removed all my existing files and folders from the old mac and want to use it as a secondary mac for a new small word processing project.


What I would like to do, is now completely erase the old back-ups from the corrupted disc and start back-ups over again for my new project.


The reason for this, is I don't trust the overwriting of old back-ups with new back-ups, as it has already proven itself as completely unreliable - and this was literally the first 'over-write' back-up.


Is there a way to 'properly' do this, or is the whole drive corrupted? It's a G-Drive 2TB - So I am also wondering, if it is not possible to erase the back-ups themselves, if there is a way to totally erase the disc - including the corrupted mac OS files and start over? Empty Mac, Empty back-up disc?


Any help appreciated, Thanks

12 replies

Bad Blocks, 'spares', SMART and DriveDx:


Since the earliest days of [only] Rotating Magnetic drives, defects in the magnetic coating might result in small bad spots. Instead of trashing an entire disc, drive-makers invented a clever work-around. They kept a small pool of spare blocks in reserve, and substituted them for blocks that were found to be bad.


When we ERASED a drive, that process would over-write every block, and read back the data to check sector integrity. Blocks found to be bad would have good spare blocks from the same general neighborhood substituted. The drive controller used a spare from a pre-ordained pool of spare blocks for each region of the drive, and kept a list of the substitutions, for the rest of its life.


Drives have gotten so big, that complete re-writing is no longer practical. That re-writing would take all day.


So today's drives 'take names' of any blocks acting up in normal operation. This whole different concept is called Self-Monitoring, Analysis, and Reporting Technology, or S.M.A.R.T. SMART was intended to help predict when a drive was going bad. Sometimes it can do that, but far from always.


DriveDx is a free program that can read out the detailed SMART status from a drive, including how many blocks are on the suspect list. That detailed information can be used in a thoughtful way to tell whether a drive is going bad, and whether it just needs a few sector repairs-by-replacement.


[Only] when new data are available for those suspect blocks, or they are to be erased, the drive controller re-writes those suspect blocks and checks those suspect blocks for sector integrity. Any found bad will get spare blocks substituted.


Punchline:

If you have the luxury to do so, ERASing the Physical DEVICE, by its immutable manufacturer-given device-name, will still get you one of two distinct outcomes. Either 100 percent good blocks, or you get "initialization failed" if not enough spares are still available in the spares pool. ("initialization failed" indicates the drive is no longer in good enough shape to Trust with your precious data.)



there are many ways data can be corrupt.


• bad blocks on the sectors holding the data

• damaged directory entries, so that pointers are no longer accurate (sometimes fixable through First Aid)

• corrupted pointers and metadata in the Time Machine data

• damaged data blocks


If you have the luxury of completely erasing the Physical device, by its immutable manufacturer-given device-name, if it passes erasure, all corruption at all levels will be eliminated.


Some of the wire-heads (like me) who are responding think there may be nothing wrong with the drive, that the data just got corrupted. (So trying to probe more deeply.)


The downside risk is that if the Physical drive fails erase, you may need a NEW drive with no advanced notice.

How did you determine they were corrupt?


Something similar happened to me once. Apple force-upgraded me to Sonoma. I restored from backup. Then git complained because it maintained its own checksums. I don't know for sure if the data was corrupt. I don't know how those checksums are calculated. But I did lose a lot of faith in Time Machine.


davey17x wrote:
Is there a way to 'properly' do this, or is the whole drive corrupted? It's a G-Drive 2TB - So I am also wondering, if it is not possible to erase the back-ups themselves, if there is a way to totally erase the disc - including the corrupted mac OS files and start over? Empty Mac, Empty back-up disc?

Yes. The only way to erase a Time Machine drive is to completely erase the disk. Erase it at the lowest possible level. You may need to use a "repartition" operation if Disk Utility refuses to erase.


My guess is that the disk is fine and the corruption came from some other source. I only say that because hardware is mass-produced in a factory by engineers. Software, however, is more of an "art form" that requires no credentials of any kind. And that was back in the good old days before AI slop.

<<. wiped all >>


If you merely erased the Mac Volume, (with name of your choosing) and did not erase the Physical Device by its immutable manufacturer-given device-name, you may have "papered over" any Physical Bad Blocks on the drive, and they may come back to bite you later.


Typing that phrase again is something I am only doing to emphasize EXACTLY what is required to remove bad sectors/blocks on the drive as you ERASE.



What does DriveDx report for the 2 TB HDD?

How did I determine they were corrupt?


I spoke to Apple support direct since I'd just bought the MBP and it didn't work, so after I'd transferred from the back-up disc, we had to wipe and reinstall (in a nutshell - there was a little bit more involved but that's the gist of it).


The corruption probably came from the 'clearing space' overwriting of the full back-up drive, as I had no problem until I transferred them to a new MBP and there were all these additional 'files' causing problems.

davey17x wrote:
How did I determine they were corrupt?

Yes. That was my question.


I spoke to Apple support direct since I'd just bought the MBP and it didn't work, so after I'd transferred from the back-up disc, we had to wipe and reinstall (in a nutshell - there was a little bit more involved but that's the gist of it).
The corruption probably came from the 'clearing space' overwriting of the full back-up drive, as I had no problem until I transferred them to a new MBP and there were all these additional 'files' causing problems.

None of that answers my question. I guess you're referring to "it didn't work"? How so? I still don't think the actual data was corrupt. It's more likely that the data transferred was an invalid configuration for the new computer.


I've never heard of DriveDx.

It's a tool for reading the SMART data from a hard drive. The data is rarely useful, but it can't hurt.

As my post says: 'How did I determine they were corrupt? I spoke to Apple support direct...'


I'm not sure why that sentence doesn't answer your question. Several paragraphs detailing the whole exchange wouldn't add anything, but I was being led by experts who determined a portion of extraneous files were making mischief.

Thanks for your contribution. TBH I don't know exactly what the corruption was, (a bit above my pay grade) except that in the overwriting process of the full back-up drive, there were various anomalies including 600+ 'icon' files which got transferred with the rest of my files to my new MBP, they didn't belong anywhere or to anything and were making the new macbook do weird things or nothing at all.


So, after trying various fixes and reboots, we basically wiped all and started over again.

Grant Bennet-Alder wrote:
…
If you have the luxury of completely erasing the Physical device, by its immutable manufacturer-given device-name, if it passes erasure, all corruption at all levels will be eliminated.

Though an HDD failing will collect new corruptions.

Some of the wire-heads (like me) who are responding think there may be nothing wrong with the drive, that the data just got corrupted. (So trying to probe more deeply.)

I’m not a big fan of erasure as a resolution for hardware-caused corruptions, particularly for HDDs that are collecting errors. It defers the inevitable, at extra effort and extra risk.

The downside risk is that if the Physical drive fails erase, you may need a NEW drive with no advanced notice.

Yep. Which is why a peek at the diagnostic and failure data can help decide.

That's like asking what does my mum report. I've never heard of DriveDx. 😊