Library and / or Application Support redirect

This used to work (sort of) back in the Xserve days running OS X server, although it was a bit slow. It's been pretty badly broken in the past 5-6 years.

In some specific cases we sync some user data to the network drive (for the purposes of backups), but as far as network based home directories go, it's not something I would recommend.

Yeah, I'm aware of the issues over the years. We had a really stable setup with nethomes last year so I was hoping to continue that without disrupting everyone's workflows.

Do you run a shared use lab environment? What's your experience with best practice for stability & file backup/transfer between stations?

We are an MSP supporting lots of schools so have a bit of a mixture. The most stable setup is network logins with local home folders and network homes mapped onto the users desktop.

Users are advised that data on the network drives gets backed up. Data on the local machines does not. This works very well and is generally accepted in most cases. The side benefit is that the students learn how to manage their files and work with network drives, and are in full control over where their data is, which is often welcomed.

In a few schools we have other processes added on, either home syncing (a custom solution rather than the OS X one) or some kind of device backup software. In these cases the complexity and cost does increase considerably.

THanks David! That makes sense.

How do you manage those local accounts not getting out of control with data (I envision our students storing way to much on them and suddenly seeing full HD errors across campus) This may be a campus-specific problem as we are an art college relying heavily on Adobe programs (in the past we've had them all work off HD partitions adequately labelled 'working volume' and moving files back onto their network account one they've been saved)

In some cases we set the /Users folder to empty each day, so we're really saying to the users "once you're done, put your work on the server, or it will be gone tomorrow". It's a bit tough, but a very stable setup.

The other method is to have a smart group identifying hard drives that are getting too full, with a policy to empty them.

You can apply a lot more logic to it with a script using -mtime to find files that haven't been modified for a while and just deleting those.

Cool. I think we'll follow the path of the smart group. Thanks for helping me wrap my head around this more concretely.

Do your users change their passwords via the Macs, or do you have a password portal (I can see the keychain errors now; will be working on an easy fix for the end user, might be time to implement ADPassMon) just curious how you handle that generally. -no rush; closing time for the week-

@agrosvenor if you delete th keychains, then they'll not have keychain unlock issues as a new keychain is created at login.

This does mean that folks cannot store passwords in the keychain though.

If the Macs are AD bound, they can change via their Macs or AD & when they login the Macs will auth to AD & create a login keychain with the correct password that lets them login.

@agrosvenor If you are still keen on redirecting specific folders from the network home to the local machine, I've been using the com.apple.MCXRedirector options with great success. You can create a Configuration Profile and add a Custom Setting payload, then upload a plist with the required settings. Here is a an example of someone using it to redirect the whole ~/Library/Application Support/ folder, though I'd probably just redirect the Chrome specific one.

I am currently redirecting the following folders for my network home users to the local disk:
~/Library/Accounts
~/Library/Safari/Databases
~/Library/Logs
~/Library/Caches
~/Library/Saved Application State

We've begrudgingly forgone the days of network homes for network login with local homes. It's been a functionality fight the past few releases, and seems to only be getting worse.

I do appreciate your note; tempting to try since someone else is still holding on to the network homes; but we've begin the process of workflow retraining.

Thanks!