Orbit fails to start on some hosts

[roperzh]

Fleet version: fleetd 1.17.0

Web browser and operating system:

---

💥  Actual behavior

Spotted at a customer env, they have an agent that sits alongside fleetd. They noticed that some hosts (let's say 5% of the total) were offline in Fleet, but reported as online by the other agent.

Upon digging we found two consistent issues across all those devices:

NSOSStatusErrorDomain

This error shows in loop, nonstop:

The application cannot be opened for an unexpected reason, error=Error Domain=NSOSStatusErrorDomain Code=-10822 \""kLSServerCommunicationErr: The server process (registration and recent items) is not available\"" UserInfo={_LSLine=507, _LSFunction=-[_LSRemoteOpenCall invokeWithError:]}\nThe application cannot be opened for an unexpected reason

🧑‍💻  Steps to reproduce

The error occurs when fleetd is installed while the user is not logged in. For the customer this is:

1. Log out
2. Install fleetd via Puppet (we need to find a workaround for this for our own reproduction)
3. Look at the logs for the errors
4. Login, see how the errors go away

[noahtalerman]

Heads up @roperzh there's another "Orbit crash loops" bug here: #14508

Not sure if this is related but wanted to point you to it just in case.

[roperzh]

thank you @noahtalerman!! unfortunately seems to be a different issue

[roperzh]

seems like this was already reported here and @mostlikelee found a way to reproduce in #13106 (comment)

The above error Domain=NSOSStatusErrorDomain Code=-10822 occurs when no console user is logged in. We may want orbit to backoff on launching the desktop app until an active user session is detected.

I'm working on getting confirmation if this was the case for the customer as well.

[JoStableford]

Related to a Slack conversation

[lukeheath]

Bug has aged out. Moving back to product drafting.