It's been a bit less than a week since High Sierra's release, and we've been busily updating SuperDuper! v3.0 Beta 1 to Beta 2 (which, per the usual custom, is available at the end of this post). Our thanks to all the users who took the time to download the first beta and provide feedback: it's been really helpful to have the additional coverage as we work to wrap up v3.0.

We've been really pleased with the way snapshots have been working with the new version. We haven't seen a single report of a problem due to highly active files: it's doing just what it's supposed to do, which is great. Doing it this way is going to really improve people's backup experience.

Sorry for the lack of jokes in this post. It's been a long week, and my punch-up crew is on braincation.

(Note that after a version of this post went up we found a better fix for one of the issues below; if you downloaded B2 at that time, please download and install it again.)

Format Change

A lot of people have been confused about how to format their backup drive as APFS, and are confused about how to get an HFS+ volume on the same drive as an APFS volume.

The new Disk Utility has some nice features, but they've buried a bunch of stuff in the UI. Here's how to do both.

Format the Whole Drive as APFS

  1. In Disk Utility's "View" menu, "Show all devices".

  2. Select the drive hardware, above the existing volume, in the sidebar.

  3. Click the Erase button.

  4. Choose the "GUID" partition scheme, and the plain APFS format.

  5. Erase the drive.

Add an APFS partition to an existing drive

  1. Select the external drive in the sidebar

  2. Click the Partition tab

  3. Click the "+" button below the partition diagram

  4. Size the volume as needed

  5. Choose the APFS format

  6. Click Apply

Add a new APFS volume to an existing APFS container

  1. Select an APFS volume in the sidebar that's in the container you want to add to

  2. Select "Add APFS Volume" from the Edit menu

  3. Select the options you want, including minimum and quota sizes if desired, and click Add

Now that you know how to do that, let's discuss SuperDuper-specific things.

We Go On Three?

Overall, the feedback has been quite positive, and I'm pretty happy with how v3.0 is working for most people, but (as expected - it's a beta) there are a few areas where we needed to either work around High Sierra weirdness, or fix our own bugs. Namely:

  1. Copying from APFS to HFS+ fails

    Three interesting things happened here. First, there's a special hidden folder on HFS+ called ".HFS Private Data^M" that we automatically ignore during a copy, since it's managed my HFS+ and is unique to the volume. And yes, it has a ^M character at the end of the name.

    Weirdly, when High Sierra converts an HFS+ volume to APFS, it retains this particular folder, even though it's no longer something that's needed. And, on top of that, trying to match the name, with the ^M, fails with APFS "globbling"...and thus we copy the file.

    Going APFS->APFS, this works fine, since the folder is just a folder. But going APFS->HFS+, we get an error, since it's not ignored (due to the APFS bug) and can't be written (because it shouldn't be copied).

    We've worked around this in the new update.

  2. There's no Preboot when going APFS->HFS+

    A last minute change caused us to try to bless the Preboot volume when there wasn't one, since HFS+ volumes don't have preboot. Although we continue to recommend copying APFS to APFS, as discussed in my previous post, this is now fixed.

  3. Errors when using Erase, then copy

    Alas, erase copies of APFS volumes are failing when finalizing Preboot and Recovery. This happens because of an issue discussed in the previous post: the UUID of the volume changes during the erase, but we're copying the Preboot and Recovery based on the old UUID value, so they can't be found.

    We didn't catch this because while we checked Erase with non-bootable volumes, to save time during our test pass, we stopped the bootable volume test and then checked Smart Update (since we were now going from the erased volume, and the previous erase tests passed). Dumb, and the test matrix has been updated to ensure it doesn't happen again. This will be fixed in the next beta: in the meantime, use Smart Update.

  4. Crash with disk images

    Another last minute regression (as we rolled back the image handling we didn't finish, mentioned in the last post). Fixed in the new beta.

    Remember, though - if you're copying APFS volume to images, please open/mount the image first and copy to its mounted volume with Smart Update, rather than to the file.

  5. Visible Snapshots

    We consciously show the snapshot during the beta for debugging purposes, which confused some users. Given the good results we're seeing, they're now treated as release-ready, and are hidden.

  6. Can't Create Snapshot

    Two users had an interesting problem where the system returned that snapshots couldn't be created. It looks like this happens when the source drive is being encrypted but the process isn't complete. It's a very rare issue, given the number of testers, and the amount of coverage. We're continuing our investigation, and trying to reproduce in house.

    If you're seeing this, and you haven't already contacted us, please get in touch!

  7. Resource Busy error when updating Recovery

    There are some interesting aspects to this particular problem. Based on the error, this should mean that the Recovery volume is mounted somewhere else, and we're trying to double-mount it. But we've checked for that, and it's not mounted anywhere else.

    As we were further working through the case with some users who had it happen (we've never seen the problem in house), we also saw situations where some of the files in Recovery were busy during pruning, and thus couldn't be removed.

    So, we came up with some clever workarounds which we've implemented in this version. They seem to work for our in-between-beta test users who hit the issue.

    As much as we've love to say "problem solved", we're continuing to gather data, since it doesn't make a lot of sense, while also logging additional diagnostic information in case it does happen to the broader test group.

    It's important to note that when this fails, your data has been backed up successfully, and you can restore: you just might not be able to boot from the backup. If you're stuck, as always, get in touch.

  8. Various and sundry items

    We fixed some problems caused by special characters in volume names, a few UI and log typos (no, we don't know what "Ingnorning" is either), updated the signature so it should open normally for all users, and the like.

Exit, Stage Right

So, that about does it for the Beta 2 release. Thanks again for helping out with the testing, and drop an email if you're having problems.

SuperDuper! 3.0 B2