APFS Time Machine is read-only, so how do you benchmark?

So since Time Machine has migrated to APFS we have a situation where the entire TM volume is read-only to a regular admin user. That makes sense for a lot of reasons, but it also means benchmarking a TM disk “in situ” has become difficult, say if you’re suspecting the disk is not performing so well and you want to check details.

Now TM does put a lot of information out to unified log (the “new” console log system), and that includes transfer rates, but it all pertains to TM performance, not necessarily a proxy for raw disk performance. So if you’re looking for something else that a benchmarking tool gives you, that tool likely needs write access to the TM volume, which we no longer have.

Are there any simple workarounds? I already tried removing the disk from TM, which works, but it doesn’t magically make the disk writable. Any other ideas? Obviously there’s more drastic measures (clone, remove from TM, format, benchmark, clone back, add to TM) and if you were better prepared ahead of time you could account for this as well (when setting up TM disk create second container on same disk, use one for TM and the other for benchmarking), but I was hoping there might be some little trick I could use to allow the benchmark to just run in place, without too much kerfuffle.

There is a simple workaround!

Courtesy of Howard Oakley. Just add an APFS volume to the TM disk in DiskUtilty. Run the benchmark. Get rid of the APFS volume. Done. :slight_smile:


Simon, what software do you use for benchmarking and what do you do if the results are less than expected?

Howard Oakley provides one called Stibium.

It’s up to you what you do with the results. But I’d suppose the most probable outcome is to do nothing. Purely educational.

1 Like

Yes, Stibium is excellent. And should you prefer something even simpler, there’s Blackmagic’s Disk Speed Test available on the MAS.

Indeed, benchmarking is intended to ensure you’re getting from the disk and its attachment to your system, what it should deliver. It’s also useful when you’re evaluating different setups. It’s basically part of what I would consider “making an informed decision”. Even if you’re relying on a third party for advice on what to get (say, something like Wirecutter), you want to make sure once you’ve purchased equipment and have it running as part of your setup, it’s performing to spec. That’s where a tool like Stibium comes into play.

1 Like