All Time Machine backups take too long

Following up with the resolution and information I've found.

Thanks for the BackupLoupe recommendation. It's a great tool and ultimately provided me with exactly the information I needed to figure out what was going on. Aside: I've had some trouble with BL getting stuck when indexing aborted TM backups, but the company/developer behind it is responsive and we're trying to work out the resolution of that problem.

For my TM performance problems, BL not only helped me identify old cruft that I didn't need (like Application Support files for apps I no longer use or have installed), but it also helped me understand why it was mysteriously trying to back up huge amounts of small files. Not only did this consume too much of my TM partition, but it also is what I think caused the backup to take so long. It turns out that I use the Backblaze online backup service, and that has a cache folder that contains copies of files during upload, log files, etc. Once I excluded /Library/Backblaze.bzpkg from my TM backups, performance returned to what I expected and seems to be operating well now. I confirmed with Backblaze that it's essentially okay to exclude this directory from TM. It'll cause a little extra work to recover things if I lose /Library/Backblaze.bzpkg and need to do a Backblaze restore, but it's fairly easily recovered and it doesn't compromise my backups.

The local TM backup volume was formatted as APFS (Case-sensitive). I'm deleting the volume and trying to format it just as APFS to see if that makes any difference.

Hi Tesserax, thanks for the follow up. Yes, this is an 8TB MyBook configured with a single APFS Container and (currently) three APFS Volumes. One is for a machine that will soon be decommissioned, so it'll go back to two volumes. Here's the output below.

Also, FYI. Time Machine (TM) will back up to drives formatted in either HFS+ or APFS. The latter format was designed to work best with SSDs.

Since the MyBook is a rotational drive, maybe it would be better to format it with HFS+ instead?

Here's the output requested. I renamed the machines to Remote1 and Remote2 to indicate the machines that are currently backing up over the network. Local is the machine that the MyBook is attached to, and the one that is super slow (both remote machines are actually pretty fast).

/dev/disk3 (synthesized):
   #:                       TYPE NAME                    SIZE       IDENTIFIER
   0:      APFS Container Scheme -                      +8.0 TB     disk3
                                 Physical Store disk2s1
   1:                APFS Volume ⁨Remote1 Time Machine⁩     485.8 GB   disk3s2
   2:                APFS Volume ⁨Local Time Machine⁩     1.7 TB     disk3s3
   3:                APFS Volume ⁨Remote2 Time Machine⁩   290.6 GB   disk3s1

Thanks MargeHomer. I'm backing up to a WD 8T MyBook, and the weird thing is that it used to work great. Yes, the first backup takes a long time, but after that, the incrementals go pretty quickly. The MBP laptop also backs up to a volume on the same drive over the network and that one works great. It's almost as if macOS loses any record of what it backed up before and always does a full backup.

It's still much slower than I expect and it seems like it's backing up way more data than it should be. Like multiple gigs every run, even though I know files aren't changing that often. But it does seem to be completing as when I look in my Time Machine folder, I can see multiple days that are getting backed up. I sure which TM would produce a log so I could understand exactly what is getting backed up and maybe I'd be able to exclude some more folders.

I tried to use `tmutil compare` to see what what actually being backed up. I see a few things that are suspicious, like /Volumes/.timemachine and a bunch of ~/Library/Application Support/com.apple.ap.promotedcontentd and ~/Library/Calendars/Calendar Sync Changes, and a few other things. Nothing all that big though. The thing is `tmutil compare` itself takes a very long time. It still hasn't completed. So back to at a loss for what's going on here. That said, `tmutil listbackups` is showing that backups are completing, so I dunno what to do next. I guess short of any other suggestions, I'll just live with it.

Thanks Tesserax, I'll try this. Yes, I'm using this drive only for Time Machine, so I won't put any partitions or volumes on it. I'm assuming HFS+ is what's called "Mac OS Extended (Journaled)" in Disk Utility.

I'll respond here after TM has some time to run.

Okay, none of this is helping at all. What I really don't understand is why the daily incremental backups are still trying to copy 57+GB. I'm watching `tmutil status` and it just seems like way too much is being copied.