Time Machine backup slows after Tahoe upgrade

Time Machine stopped working for me a few weeks ago.


I have three external hard drives that I use for backup and I rotate the three. After upgrading to Tahoe a few weeks ago, I successfully backed up to the first drive in my rotation. However, when I went to backup to the second, it got stuck for a very long time. It did not get stuck in the "preparing backup" phase. Instead, after copying a few gigs relatively quickly, it slowed to a crawl...we're talking a few megabytes per hour. I know that sometimes Time Machine has to delete old backups, so I let it sit...for six nights! Six nights later, it was still crawling at a few MB an hour, not having even progressed a gig in six days. I've had Time Machine get stuck for hours cleaning out old backups, but never days.


Maybe something's wrong with my second backup drive, I thought, so I'll try the third in my rotation. Same problem, so the problem wasn't the drive.


Then I decided to try a fresh backup. I reformatted a 2TB hard drive (to back up about 400GB from a 512GB machine). Same problem. Time Machine copied the first 250GB or so quickly, then slowed to a crawl. Obviously, the problem isn't that Time Machine is clearing up space on the backup drive because this was a large and empty drive.


I've tried other things...


--I restated my Mac. No difference.


--I read somewhere that you need space on the drive being backed up, not just the target drive, so I hand copied about 50GB of data off my Mac's hard drive. Even with the extra empty space on the source, the problem did not go away.


--Given that the fresh backup got through 250 GB before slowing down, I thought that maybe Time Machine is having trouble accessing something on my hard drive, so I ran Disk Utility to see if it found any problems. Other than a few minor things, it found none, and the problem continued afterwards.


I do not have any antivirus software installed. I have not changed anything on my Mac recently other than installing Tahoe, and even that isn't necessarily the culprit because, as mentioned above, I backed up once successfully after upgrading.


Any suggestions? I'm very frustrated. Thanks.





Posted on Jan 5, 2026 4:45 PM

Reply
Question marked as Top-ranking reply

Posted on Jan 7, 2026 12:47 AM

This is not like a bad drive or a Time Machine bug, it looks like Tahoe stressing APFS snapshots on a nearly full internal disk. With only -41 GB free on a 512 GB SSD, macOS is juggling local Time Machine snapshots, APFS copy-on-write blocks, and snapshot thinning while trying to stream data out, and once it hits older, more fragmented snapshot ranges, throughput can collapse to KB or MB per hour without ever saying “preparing.”


The giveaway is that every backup flies until roughly the same point, even on a fresh disk. Tahoe seems more aggressive about local snapshots than previous releases, so you’re effectively backing up a moving target with no scratch space.


Temporarily free up a lot more space, think 80–100 GB, then explicitly flush local snapshots with tmutil listlocalsnapshots / and tmutil deletelocalsnapshots <date> (or disable local snapshots briefly), reboot, and run the backup again.


If it completes normally after that, you’ve confirmed the bottleneck is snapshot churn, not hardware. Long term, keep at least 15–20% of the internal SSD truly free or Time Machine on newer macOS will keep falling off this same performance cliff.

15 replies
Question marked as Top-ranking reply

Jan 7, 2026 12:47 AM in response to TC_

This is not like a bad drive or a Time Machine bug, it looks like Tahoe stressing APFS snapshots on a nearly full internal disk. With only -41 GB free on a 512 GB SSD, macOS is juggling local Time Machine snapshots, APFS copy-on-write blocks, and snapshot thinning while trying to stream data out, and once it hits older, more fragmented snapshot ranges, throughput can collapse to KB or MB per hour without ever saying “preparing.”


The giveaway is that every backup flies until roughly the same point, even on a fresh disk. Tahoe seems more aggressive about local snapshots than previous releases, so you’re effectively backing up a moving target with no scratch space.


Temporarily free up a lot more space, think 80–100 GB, then explicitly flush local snapshots with tmutil listlocalsnapshots / and tmutil deletelocalsnapshots <date> (or disable local snapshots briefly), reboot, and run the backup again.


If it completes normally after that, you’ve confirmed the bottleneck is snapshot churn, not hardware. Long term, keep at least 15–20% of the internal SSD truly free or Time Machine on newer macOS will keep falling off this same performance cliff.

Jan 8, 2026 10:08 PM in response to TC_

This actually narrows it down.


Since freeing snapshots didn’t change the stall point, the slowdown is almost certainly happening during snapshot diff reconciliation, not space pressure anymore.


Time Machine isn’t copying bulk data at that stage, it’s walking millions of filesystem records comparing APFS snapshots, and Tahoe seems to have a nasty regression where this phase can go pathological if the backup history was created pre-upgrade. That’s why new disks, old disks, and plenty of free space all behave the same, and why the UI flips to “calculating time remaining” instead of moving bytes. At this point I’d stop trying to save the existing backup set, exclude the backup disk, reboot, re-add it, and start a brand new Time Machine backup after deleting the old TM folder entirely.

If that completes at normal speed, it confirms the issue is corrupted or incompatible snapshot metadata, not your SSD or enclosure. If it still crawls, then I’d suspect a filesystem traversal issue on the internal disk itself and test with tmutil startbackup --block from Terminal to see if it ever advances past the same inode range.


Either way, letting it sit won’t resolve this, it’s stuck doing bookkeeping, not waiting on I/O

Jan 14, 2026 1:04 AM in response to TC_

For reference: If Time Machine is slow on Mac - Apple Support



At your option use the following shell script to extract TM's recent log entries:

clear; printf '\e[3J' && log show --predicate 'subsystem == "com.apple.TimeMachine"' --info --last 2h | awk -F']' '{print substr($0,1,19), $NF}'


Copy that entire line of text (triple-click to select the entire line) and paste it into a Terminal window. It will report the previous 2 hours of activity. To change that period of time change its --last value from 2h to something else. For example you can use 24h or 30m for 24 hours or 30 minutes respectively.


Although you can run that script at any time and as often as you wish I suggest waiting for a backup to be in progress to capture the events leading up to that time. Running it an hour or two after it appears to stall will reveal what it has been doing during that period of time.


Then, Copy that that entire Terminal window's text (⌘ a ⌘ c) and post it to this site using the "additional text" icon shown below:



Needless to say, if you are using any non-Apple "anti-virus", "cleaning", or "Internet security" products including "banking security", so-called "doctors" / "cleaners" / "optimizers" and so forth, anything at all in that broad category of utterly useless unmitigated garbage, don't. A lot of things won't work if you do.


Still need help? Consider using EtreCheck. Instructions: How to use the Add Text Feature When Posting Large Amounts of Text, i.e. an Etrecheck Report - Apple Community

Jan 14, 2026 9:42 AM in response to TC_

The failures are definitely related to local snapshot storage. Of course that shouldn't be happening.


The first and most fundamental step is to restart the Mac, but you did that already to no apparent effect. Next step would be to address those Local Snapshots but first please download and run the EtreCheck report, and post its report in a reply.


If you want to "read ahead" about addressing the Local Snapshots refer to About Time Machine local snapshots - Apple Support. Step 3 instructs us to wait "a few minutes" but in my experience I'd give it longer than that. An hour or two is reasonable. You might even want to restart the Mac. Probably not necessary but why not.


Another thing I noticed is "Ignoring error 100092 on item which is unreadable while device locked: '/Volumes/com.apple.TimeMachine.localsnapshots/Backups.backupdb/Tom’s MacBook Air M1/2026-01-14-073222/Data/Users/tchristy/Library/Containers/47A73F24-D77C-4A77-95DC-A9CE9439EC4D/Data/Library/Application Support/CouchbaseLite/webservice_db.cblite-shm'", and various other files in the CouchbaseLite folder. I never installed anything called Couchbase, though perhaps one of my apps did.


I don't know what it is either. I would be inclined to drag that Application Support folder to the trash, if not for the fact I have no idea if it's required by something you installed. In the interest of taking a conservative approach (and the fact your backups are questionable) leave it be for now. The EtreCheck report may reveal some additional information.

Jan 5, 2026 6:58 PM in response to TC_

TC_ wrote:
--I read somewhere that you need space on the drive being backed up, not just the target drive, so I hand copied about 50GB of data off my Mac's hard drive. Even with the extra empty space on the source, the problem did not go away.

How much free space is on your internal drive? Not what macOS reports as 'available' in Get Info on the drive or System Settings > General > Storage, but what is reported by Disk Utility or in System Settings > General > About > System Report.

Jan 14, 2026 8:48 AM in response to John Galt

Thanks.


I have no anti-virus, cleaning, or non-Apple security products installed.


First, an update: after the brand new backup got stuck last night, I stopped it, then re-booted in Safe Mode. It successfully finished the backup from Safe Mode after copying another 4 GB. Looking through the directory on the backup drive, I can see files that I worked on yesterday, so I believe that everything is there.


This morning, I tried running it again from regular mode. It has copied about 107GB so far (even though not that much could have changed since last night) and then slowed down, claiming that over 22 hours remain (which number keeps increasing).


I ran the log command in Terminal with it still running. I'm not going to attach the whole log because it's very long and very repetitive. I am just copying through where it says "starting backup pass 2," after which things become repetitive.


I'm guessing that the problems are related to the errors "permission denied," "structure missing," and/or "structure has wrong type."


Another thing I noticed is "Ignoring error 100092 on item which is unreadable while device locked: '/Volumes/com.apple.TimeMachine.localsnapshots/Backups.backupdb/Tom’s MacBook Air M1/2026-01-14-073222/Data/Users/tchristy/Library/Containers/47A73F24-D77C-4A77-95DC-A9CE9439EC4D/Data/Library/Application Support/CouchbaseLite/webservice_db.cblite-shm'", and various other files in the CouchbaseLite folder. I never installed anything called Couchbase, though perhaps one of my apps did. Googling around, I couldn't find too much information about it. However, because TM says that it's ignoring the error, I'm guessing that's not the problem.


Finally, I noticed a message "Prioritizing backing up 126 items which are inaccessible while device is locked." I'm not sure if that has anything to do with things.


All help appreciated.









Jan 8, 2026 9:01 PM in response to iamshivam

Thanks for the suggestion. I deleted another 25GB or so from the local drive in addition to the 27GB that I had already deleted. Interestingly, it made no difference in the amount free reported in System Information, which stayed stuck at 30some GB.


I then deleted the local snapshots, which were taking about 90GB. That not only freed up 90GB, but also seemed to make the system recognize the earlier deletions. All of a sudden, I had over 150GB free.


So far, however, it has not helped with the backup. I am currently backing up to one of my existing backups. Two hours in and I seem to be stuck at the same point. The only difference is that, in System Settings, where you can get detailed information about the backup progress, instead of reporting an estimated time (which was constantly increasing), now it says "calculating time remaining" and seems stuck on that message. I will let it sit overnight and see what happens, but no luck yet despite having about 30%+ of the drive free.




Jan 13, 2026 8:03 PM in response to iamshivam

Thanks. Still experiencing some frustration here...


As I mentioned in my original post, when I first tried a fresh backup, it got to about 250GB before getting stuck. Then, as mentioned later in the thread, I succeeded with a fresh backup from safe mode. As you suggested, I decided to give up on my old backup sets and tried another new set (on a brand new drive). I did not run from safe mode. It got to 268GB out of about 400GB, reporting 70% or whatever it was copied, then suddenly got stuck, still reporting 268GB copied, but now calling it 0.0% and not going anywhere. This is essentially the same point where it got stuck as mentioned in my original post.


If I try your tmutil startbackup --block suggestion, what am I looking for? Should I run that from safe mode?


Thanks. I'm glad that someone around here knows how TM actually works...


P.S./edit: I know that most Apple stuff no longer logs to the old Console app, and I never learned how to search logs in the new unified logging, but is there anything worth looking for there?

This thread has been closed by the system or the community team. You may vote for any posts you find helpful, or search the Community for additional answers.

Time Machine backup slows after Tahoe upgrade

Welcome to Apple Support Community
A forum where Apple customers help each other with their products. Get started with your Apple Account.