I started seeing TestAPISwarmLeaderElection fail frequently in CI. Searching through issue history, this seems to be new

Here is an example failure: https://jenkins.dockerproject.org/job/Docker-PRs-powerpc/1828/console

Logs from that failure: https://gist.github.com/aaronlehmann/c46e85d7626431e9c4dfe857f77f9df8

I can't think of any recent changes which would cause a node to trigger a leader election on restart. One possibility is that #32283 changed the timing of shutting down the node, and this somehow has an impact.

I don't think there's anything bad about a node triggering a leader election when it starts up, and I don't know that we can provide a guarantee that this won't happen, so it may make sense to change the test to tolerate this condition.

cc @aboch