Tahoe Free Space Problems
Running the Mac OS 26 beta (I know, I know), but running out of hard drive space. Turns out the reason is that half of my drive is full of “System Data”.
Internet searches are not very helpful in terms of how to reclaim this space. I’ve tried booting into safe mode and back, but wherever these files are isn’t visible to anything like DaisyDisk to check further and/or persuade them to go away.
I had a similar issue on my beta 5, updating to beta 6 didn’t fix anything. Additionally, there seemed to be a memory leak within the mds_stores process, which was using a lot of my memory and it was increasing every second up until it started using my swap.
I couldn’t find a solution, so I backed up my Mac (created a clone using SuperDuper!), clean installed beta 6 again and restored my backup. Issue is gone, reclaimed over 100 GB of free space and System Data now shows around 26 GB.
I had this with beta 5. It was crazy watching the disk space tick away for no reason. Beta 6 fixed it, at least so far it has.
I’ve had lots of problems with Tahoe operations (primarily xcodebuild
) failing because it reported no free space, and this continued with beta 6. Whenever this happened, Finder and Disk Utility would both show about 40 GB of free space. There did not seem to be any hidden usage due to snapshots, which makes sense since I’m not backing up that Mac installation. I have no idea what else would be using the space, why it doesn’t see to show up as used anywhere, nor why it eventually seems to clear up on its own, but only for a time.
Previously:
- macOS Tahoe 26 Developer Beta 6
- macOS Tahoe 26 Developer Beta 5
- Spotlight Indexing Running Wild
- Fixing “Optimize Storage”
- Clearing Space on Your Mac
- Pruning iOS “System Data”
- Quantum Computing and APFS: Free and Used Space
1 Comment RSS · Twitter · Mastodon
My MacBook Pro went wild last week, once I’ve updated to Tahoe Beta 6. Looks like Spotlight was the culprit, showing high CPU usage and byte writes on `mds_stores` and the entire disk space being eaten.
This is not a new bug, as per your reference article. Users on Sequoia have experienced similar problems. Maybe a regression?