I’m running TrueNAS Scale 24.10 and have a dataset shared via SMB using the Basic Time Machine Share preset. I back up my Mac to this share once a week, and it typically uploads around 10 GB of new data.
After each backup, I have a Cloud Sync task to sync the dataset to a Backblaze B2 bucket. However, during the dry run, it reports over 150 GB of changed data to upload—even though only ~10 GB was added by Time Machine. It does this every week.
This is a problem because I have a 1 TB monthly upload data cap, and these unnecessary uploads are eating into it fast.
Has anyone dealt with this issue? Is there a way to reduce the amount of data that gets uploaded to B2 while still using Time Machine?
Time Machine is a disk image split into bands. When you create a new backup new data is written and it may result in multiple bands added and/or modified. Then when you sync each band needs to be uploaded in its entirety, even if only a few bytes changed. If your Time Machine is encrypted — the data spread is even more amplified.
The advice is to not sync Time Machine to Backblaze. Time Machine is a local backup. It contains a lot of data that is nice to have if bandwidth and storage is not a concern but can be 100% omitted from offsite backup. Stuff like derivative data in your photos library and browser databases.
Instead, use another tool to backup your actual source data (without derivative; using photos library again as an example — pick only originals folder) to Backblaze. There are many options to choose from. On a Mac Arq7 is a solid choice. It will also take care of dataless files, which are missing from Time Machine but still need to be backed up.
I also use Arq on Mac (as well as Windows) and have for almost ten years. It’s never let me down and it’s one of the first five applications I install on a new non-Linux system. Dataless file materialization is a fantastic mechanism to backup cloud-synced directories. No one ever remembers that, but I have yet to meet a cloud provider whose default terms include responsibility/commitments for backups. If you’re on a metered Internet connection this feature can be an unpleasant surprise depending on how much data you have in the cloud directories you have in-scope of the backup set. I recently had a friend using Arq run into that issue on their Comcast connection.
I like the suggestion to keep just original photo files in an off-site when bandwidth is a concern. I’d just caution that the data that won’t be backed up would include (but not be limited to) recognized faces/pets, saved edits, comments, album, tagging, and iMessage cross-reference data. If you’re trying to ensure you have a full off site backup, then backing up only originals may or may not be what you want depending on what you define as a “full” offsite backup.
Worth noting, it will materialize a file only once, unless it changes; therefore only the initial backup will consume a lot of bandwidth – but Comcast forgives you for blowing though the bandwidth allocation once (or twice?)
Speaking of Comcast – they like to play games where they overcharge existing customers while providing long and juicy “deals’” for new customers. They want to play these games – so I’m playing these games. Now I move in into my house as a new customer every time the upcharge begins “promotional” rate expires. Recently they offered 5 years price guarantee for new customers for unlimited service plus free equipment rental at the half the cost of what loyal customers pay for metered connection. So I moved in into my house again
The point being – on Comcast one can easily get unlimited data, and get a discount, by playing their games, and therefore not sacrifice backup quality in fear of being overcharged for traffic.
So true. Most people don’t read terms of the services they use. For example, this is form iCloud ToS:
Apple shall use reasonable skill and due care in providing the Service, but, TO THE GREATEST EXTENT PERMISSIBLE BY APPLICABLE LAW, APPLE DOES NOT GUARANTEE OR WARRANT THAT ANY CONTENT YOU MAY STORE OR ACCESS THROUGH THE SERVICE WILL NOT BE SUBJECT TO INADVERTENT DAMAGE, CORRUPTION, LOSS, OR REMOVAL IN ACCORDANCE WITH THE TERMS OF THIS AGREEMENT, AND APPLE SHALL NOT BE RESPONSIBLE SHOULD SUCH DAMAGE, CORRUPTION, LOSS, OR REMOVAL OCCUR. It is your responsibility to maintain appropriate alternate backup of your information and data
The last sentence is pretty much the statement in plaintext.
That’s fair – I guess it’s a personal decision of what to consider “derivative” data. Some closer review of the photos library can reveal truly derivative locations – for example, generated thumbnails, and converted jpeg files, that can definitely be excluded. So the approach can be to grab everything and then exclude what’s definitely not needed. There is a number of databases inside that may or may not be important. Another question is whether those can be restorable – that woudl need to be tested.
