macOS setup experience: end users are stuck for over 1 hr

[zayhanlon]

Fleet version: 4.66.0 RC

---

💥  Actual behavior

This is the most critical phase in the onboarding process. Customer walked us through what he was seeing on a recorded video: https://us-65885.app.gong.io/call?id=4210680927126709117

1. After an hour plus of the software attempting to install

2. The progress spinners stop and everything kind of freezes

3. No error messages or timeouts

4. Nothing exposed to the end user or Fleet admin's team

5. No one knows what happened here

6. After reboot, able to bypass the setup experience and allows access to the device

7. If any items fail, we don't want to release the device to the user

8. Dialogue comes back

9. Still trying to install

10. But it clears past the software install page

11. No guard rails for the user when they’re brought into the experience### 🧑‍💻  Steps to reproduce

12. In Fleet, under Controls > Setup experience, add software

13. Open a freshly wiped Mac and wait for Fleet's Setup experience window

14. When the window appears turn off your local Wi-Fi to simulate slow/no internet connection.

@allenhouchins: Turning off Wi-Fi might actually cause Fleet to show "Failed" and the let the end user through.

🛠️ To fix

@marko-lisica: TODO

[noahtalerman]

@zayhanlon we flipped this request to a bug. I assigned you. Can you please help us add the Fleet version and the video the customer shared?

After that, please add back the :product label so that we can get moving on the fix!

[zayhanlon]

We're taking it on the agenda for tomorrows meeting. Version is 4.66 RC (updated above). Thanks! @noahtalerman

[sharon-fdm]

@noahtalerman , @zayhanlon , I'm assuming this is a released bug and not just in 4.66RC, because it's reported by customers. Please tell me if you think I'm wrong.

cc @xpkoala, @lukeheath

[zayhanlon]

@sharon-fdm yes we know that they saw the same in 4.65.0 as well, but if you want we can wait for the 4.66.0 upgrade this week and check back again

[lukeheath]

@zayhanlon It sounds like they are already on the RC? If so, trying again on 4.66.0 won't address it unless we get a bug fix in today.

[zayhanlon]

@noahtalerman Gong attached up top! Thanks @marko-lisica fyi

[zayhanlon]

@noahtalerman can i add p2 to this product design issue? it's workflow blocking for macOS mdm migration

[noahtalerman]

@zayhanlon this is a top priority for the #g-mdm in the upcoming design sprint.

Reminder to assign @lukeheath (DRI for what goes in a release) to issues when you add priority label: https://fleetdm.com/handbook/company/communications#high-priority-user-stories-and-bugs:~:text=Comment%20on%20the%20GitHub%20issue%20explaining%20why%20the%20issue%20is%20a%20priority%2C%20and%20assign%20the%20issue%20to%20the%20DRI%20for%20what%20goes%20in%20a%20release

[zayhanlon]

@noahtalerman thank you! yes, wasn't sure if it was the same process for things going to design. thanks again!

[lukeheath]

Thanks, y'all. P2 looks appropriate to me.

[nonpunctual]

https://github.com/fleetdm/fleet/issues/27820

[noahtalerman]

(more info in the Google doc here):

• #28434
• #28450
• #27820
• #28448
• #28451
• #28452
• #28205

FYI @zayhanlon, as of 4.69, we addressed all but one of the items.

This is the one: https://github.com/fleetdm/fleet/issues/28451

• We think we misunderstood the current behavior: Fleet doesn't retry installing software.
• It sounds like, in this scenario, what the end user sees is "Failed" and the IT admin, in the logs, might see Fleet's agent (fleetd) retrying to send the install results (install script output) to Fleet.
• The software will stay "Failed" because Fleet doesn't retry installing the software.

We think this is acceptable behavior but it's up to the customer. Can you please let them know and ask for their feedback on this last item and all other items?

I added this bug to the :help-customers board. Up to you to close or leave open with feedback: https://fleetdm.com/handbook/customer-success#communicate-feedback-on-prioritized-customer-requests

[AndreyKizimenko]

Ran into it myself on 06/08 during the onboarding while setting up a new machine.

1. Was initially stuck on the "Waiting for management server" and "Connecting to server: dogfood...".
2. Waited for 2-3 hours and had to hard shutdown to see if that helps
3. On restart I was able to bypass all the enrollments and setup a new user
4. On the next restart my machine was locked out. First it went to "Setup user" screen but after another restart it remained in "Welcome to Mac" with any action restarting the laptop.

My solution was a full wipe and after that dogfood server properly picked up my machine within a couple of minutes.

[zayhanlon]

Hasn't come up again on new enrollments. Will reopen if needed

[fleet-release]

Software setup stalls, Fleet now ensures smoother sail, Ease blooms, like spring's thaw.