Posted by on October 17, 2019

After upgrading to macOS Catalina 10.15, your previously working autofs mountpoints may need to be updated with Catalina’s new filesystem structure.

In this example, I was previously mounting NFS shares from my local NAS to a folder named “Nas” in /Users/me. In Catalina, your data moves to a separate volume from the operating system files. (You can verify this in Disk Utility.)

New “Data” mount point in Disk Utiility – Catalina 10.15

So from the perspective of autofs, your previous mount points may no longer exist at the previous location. In my case, the /Users directory moves to /System/Volumes/Data/Users.

To fix my /etc/auto_nas file, I simply prepended each mount point with /System/Volumes/Data:

# Media share
/System/Volumes/Data/Users/me/Nas/media     -fstype=nfs,noowners,noresvport,hard,bg,intr,rw,tcp,rdirplus,rsize=65536,wsize=65536,readahead=128 nfs://my-nas:/mnt/Pool1/media

# Storage share
/System/Volumes/Data/Users/me/Nas/storage     -fstype=nfs,noowners,noresvport,hard,bg,intr,rw,tcp,rdirplus,rsize=65536,wsize=65536,readahead=128 nfs://my-nas:/mnt/Pool1/storage

# Archives share
/System/Volumes/Data/Users/me/Nas/archives     -fstype=nfs,noowners,noresvport,hard,bg,intr,rw,tcp,rdirplus,rsize=65536,wsize=65536,readahead=128 nfs://my-nas:/mnt/Pool2/archives

With your mountpoints corrected, save the file, run the automount command (to flush the cache and reload /etc/auto_master), and get back to enjoying the new OS!

sudo automount -vc

If automount completes successfully, but does not confirm that it mounted your shares, you may need to update your automount master mapping file (/etc/auto_master) as well.

UPDATE (2020-02-01)

A word of caution about using a mount point in your Users directory

While automount points inside /System/Volumes/Data/Users/<you>/ automount successfully, you may encounter a peculiar issue in Catalina where the mappings duplicate over time (seemingly each time they get re-mounted by autofs), and CPU utilization increases to the point of system instability. If you run the mount command, it will report any duplicates, even though you will only see one instance of the mount point in Finder. When the duplication occurs, the automountd process will consume significantly more CPU cycles than it normally would.

The solution is frustratingly simple: move your automounts outside of the /System/Volumes/Data/Users/ directory. After struggling with this for months, I moved my mount points to /System/Volumes/Data/Nas and everything went back to normal (or at least, to how it behaved pre-Catalina).

Relevant threads from discussions.apple.com:

Comments

  1. Graham Heathcote
    March 26, 2020

    Leave a Reply

    That last update to move the mounts out of User directory seems to have fixed my problems. Thanks!

  2. Forrest
    April 20, 2020

    Leave a Reply

    In this context, with Catalina, how would we reliably use NFS-mounted HOME directories instead of the local mount point?

  3. Kyle
    April 20, 2020

    Leave a Reply

    I know what you’re referring to, but I haven’t tried that with Catalina. Hopefully you can find some documentation from Apple if they still support it.

  4. Luis
    July 4, 2020

    Leave a Reply

    Hello Kyle,

    Thanks for this post! I finally found a solution to my nfs mount problem. I tried your suggestion to mount my nfs share outside of my User directory but although automount says that it has mounted the share and I can also see it when I use df, the share is not really mounted. That is, I cannot cd to the directory. It’s basically just empty. df shows

    map auto_nfs 0 0 0 100% 0 0 100% /System/Volumes/Data/test5

    unlike when I mount it inside my User directory and I can see the source of the nfs share.

    I tried creating a mountpoint first and also letting auto_nfs create the directory but it still returns empty.

    Is there anything else i need to do to create the mountpoint outside of my User directory?

    Thanks a million in advance! I am ready to go crazy! 🙂

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: