[Bug]: MacOS 15.0.X TimeMachine Backup is skipping #185
Comments
mbentley
changed the title
|
The "multiple responses" line is because your WORKGROUP is apparently not unique on the network. You can try to set it to something besides WORKGROUP but I am not sure if that will change anything. It really doesn't matter at all for TimeMachine usage and can be a random string for all it cares. I know it's not really helpful but I haven't had any specific issues with macOS 15 and Time Machine. My two hosts running 15 have backed up without any issues or config changes required. I haven't yet gone through all of the configuration options you have in your container vs the default yet but just thought I would at least share the bit about the workgroup for now. *edit: I am really only seeing two defaults that are different from the default of: SMB_NFS_ACES: "no" I will have to find the issues where I've changed those but I also don't believe that these would specifically cause the issues you're seeing with the share disconnecting. |
|
I also noticed the message with the multiple responses and initially dismissed it as a cause of the problem since it still works without issues on 14.X. I have now changed the Workgroup name to a unique one, but the problem persists. I also removed the non-standard configurations, but unfortunately, this didn’t help. In fact, without these settings, the backup fails almost immediately. The fact that the backups are running smoothly for you is interesting. Which server operating system are you running the Docker instance on? Do you happen to have any other ideas on what could be causing this? I’m a bit at a loss here. The following areas seem to be running fine: I created a new server (Ubuntu) and used different drives, but the problem persists. To me, this narrows the issue down to the network and the Mac devices themselves. Since it works without issue on 14.X, I suspect the problem lies with the Mac. I would rule out WLAN issues, as it also occurs with LAN connections. This again leads to the assumption that Apple may have made some adjustments to SMB, causing these problems. However, what puzzles me is that it works fine for you. If you have any further ideas, I would be grateful. |
|
I guess I'm not surprised that the workgroup change didn't impact anything. Changing the settings for the backups seems like it causes them to fail unless you re-create them in some cases - not exactly sure why but I recall seeing similar behavior in the past. I'm using Debian 12 (bookworm) for the Docker host OS. My persistent storage for time machine is on a zfs filesystem and I am using macvlan for my container's networking. My run command is very basic: docker run -d \
--restart unless-stopped \
--name timemachine \
--hostname timemachine \
--network macvlan1 \
--ip 192.168.2.162 \
-e DEBUG_LEVEL="1" \
-e MIMIC_MODEL="TimeCapsule8,119" \
-e PASSWORD="password" \
-e SET_PERMISSIONS="true" \
-e SHARE_NAME="TimeMachine" \
-e VOLUME_SIZE_LIMIT="3 T" \
-e WORKGROUP="WORKGROUP" \
-e IGNORE_DOS_ATTRIBUTES="true" \
--mount type=bind,source=/zfs/storage/timemachine/data,destination=/opt/timemachine,readonly=false \
--mount type=bind,source=/zfs/storage/timemachine/logs,destination=/var/log,readonly=false \
--mount type=tmpfs,destination=/run/samba \
--cpus 4 \
--memory 1g \
--memory-swap 2g \
mbentley/timemachineOne thing I might suggest is checking out TheTimeMachineMechanic (T2M2) from here. It tries to pull together logs from the Mac side of things to determine why a backup might be failing. |
|
I'm having this same issue on macOS 15.0.1. This container is running on a mini PC home server running the latest version of Debian. I just tried deleting my backup and setting up Time Machine again on my Mac but it's still having the same issue. My docker-compose.yml: samba:
image: mbentley/timemachine:smb
container_name: samba
restart: unless-stopped
environment:
- PASSWORD_FILE=/run/secrets/samba_password
- SHARE_NAME=Backup
secrets:
- samba_password
volumes:
- /home/myuser/backup/timemachine:/opt/timemachine
tmpfs:
- /run/samba
ports:
- 172.16.16.1:137:137/udp
- 172.16.16.1:138:138/udp
- 172.16.16.1:139:139
- 172.16.16.1:445:445I don't know if it's related, but I did notice in the container logs some error messages that I don't remember seeing before: |
|
I too has same issue after updating to 15.0.1. Tried to install samba on host machine (Ubuntu 24.04.1 LTS), reinitialize backup, change from encrypted to non - same behavior. I do also see My compose config: (healthcheck.sh is similar to one in the repo, but slightly changed to smbd, container is automatically updated with new versions via watchtower) |
|
I would suggest checking out TheTimeMachineMechanic (T2M2) from here to see if there is anything showing up in your Mac's logs that would indicate anything client side to help try to narrow that down. |
|
Here is an excerpt from T2M2: In addition, I have done the following things: The permissions issue is a bit confusing here. As tested, the storage location can be accessed and edited without any problems. I don’t quite understand why T2M2 and the internal SAMBA logs are throwing errors. |
|
Maybe try just changing |
Isn't it a default value, as per Entrypoint? I've looked at logs of T2M2 on my machine and saw |
Sorry, I should have specifically called out @exomor for that specific value because yes you're correct that is the default setting. It's really difficult to figure out what in the logs is really a warning or if they're actual errors as there are many messages that sound like critical failures but turn out to not be. That's interesting about the firewall. Makes me wonder if there is something with the firewall resetting states and re-applying the firewall rules or something like that. I do have the macOS firewall enabled on my Macbook but it would be interesting to see if anyone else can try to temporarily disable the firewall and try to take a backup and see if it is successful. |
Unfortunately, the error persists. This option doesn’t seem to be causing the trouble.
That’s it. The firewall is causing the problem. I currently have two backups running. My MacBook Air successfully completed the last backup. The first backup from my MacBook Pro is still in progress. It will take a little longer, but after ~120GB, I’m feeling pretty confident. edit: My primary backup was also completed successfully edit edit 17.10: |
|
Just small FYI: 15.1 released and issue was fixed on Apple's side without any mention in release notes |
and others
Describe the Bug
Hello mbentley,
thanks for your very helpful docker container! I'm using it a lot!
Unfortunately since the MacOS Update to 15.0.X there are on multiple clients some issues.
When the TimeMachine backup starts, the client finds and connects to the container. It also detects that there are changes and begins copying them. However, after a short time (a few seconds to 2 minutes), there are repeated interruptions with the error message “The backup disk image was ejected or disconnected from your Mac while backing up.”
This issue occurs over both WLAN and LAN. On a test device with macOS version 14.X, these problems do not occur. However, the problem exists on all clients with version 15.0.X
Expected Behavior
That the connection dont fail.
Steps to Reproduce
How You're Launching the Container
Container Logs
Time Machine client Logs
Additional host information
Additional Context
In the Docker logs, you can see a message indicating “(query_name_response: Multiple responses…)” multiple responses. However, I have verified several times that 192.168.X.60 is used exclusively by the Docker container.
I had to shorten the Logs because of the character limit. I added the most important log parts.
I'm assuming that Apple has changed the way how it uses SMB. So I don't think that further updates of MacOS will solve the communication problem with the docker container. What so you think about it?
I would appreciate any help
The text was updated successfully, but these errors were encountered: