mirror of
https://github.com/juanfont/headscale.git
synced 2026-10-27 08:46:49 +09:00
The expirable LRU ran its sweeper forever, one goroutine per State. Purge first so pending waiters get ErrRegistrationExpired instead of hanging. The OIDC state cache is closed at server shutdown for the same reason.
154 lines
5.1 KiB
Go
154 lines
5.1 KiB
Go
package servertest_test
|
|
|
|
import (
|
|
"testing"
|
|
"time"
|
|
|
|
"github.com/juanfont/headscale/hscontrol/servertest"
|
|
"github.com/stretchr/testify/assert"
|
|
"github.com/stretchr/testify/require"
|
|
"tailscale.com/types/netmap"
|
|
)
|
|
|
|
// TestEphemeralNodes tests the lifecycle of ephemeral nodes,
|
|
// which should be automatically cleaned up when they disconnect.
|
|
//
|
|
// TODO(kradalby): These tests wait for real-time grace periods and
|
|
// GC intervals (up to 60s). testing/synctest would allow instant
|
|
// fake-clock advancement, but two blockers prevent adoption
|
|
// as of Go 1.26:
|
|
//
|
|
// 1. database/sql internal goroutines: Uses [sync.RWMutex] which is not
|
|
// durably blocking in synctest, causing hangs.
|
|
// - https://github.com/golang/go/issues/77687 (mutex as durably blocking)
|
|
//
|
|
// 2. net/http server goroutines: I/O-blocked goroutines are not durably
|
|
// blocking, preventing bubble termination.
|
|
// - https://github.com/golang/go/issues/76608 (httptest synctest support)
|
|
func TestEphemeralNodes(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
t.Run("ephemeral_connects_and_sees_peers", func(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
srv := servertest.NewServer(t,
|
|
servertest.WithEphemeralTimeout(5*time.Second))
|
|
user := srv.CreateUser(t, "eph-user")
|
|
|
|
regular := servertest.NewClient(t, srv, "eph-regular",
|
|
servertest.WithUser(user))
|
|
ephemeral := servertest.NewClient(t, srv, "eph-ephemeral",
|
|
servertest.WithUser(user), servertest.WithEphemeral())
|
|
|
|
// Both should see each other.
|
|
regular.WaitForPeers(t, 1, 10*time.Second)
|
|
ephemeral.WaitForPeers(t, 1, 10*time.Second)
|
|
|
|
_, found := regular.PeerByName("eph-ephemeral")
|
|
assert.True(t, found, "regular should see ephemeral peer")
|
|
|
|
_, found = ephemeral.PeerByName("eph-regular")
|
|
assert.True(t, found, "ephemeral should see regular peer")
|
|
})
|
|
|
|
t.Run("ephemeral_cleanup_after_disconnect", func(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
// Use a short ephemeral timeout so the test doesn't take long.
|
|
srv := servertest.NewServer(t,
|
|
servertest.WithEphemeralTimeout(3*time.Second))
|
|
user := srv.CreateUser(t, "eph-cleanup-user")
|
|
|
|
regular := servertest.NewClient(t, srv, "eph-cleanup-regular",
|
|
servertest.WithUser(user))
|
|
ephemeral := servertest.NewClient(t, srv, "eph-cleanup-ephemeral",
|
|
servertest.WithUser(user), servertest.WithEphemeral())
|
|
|
|
regular.WaitForPeers(t, 1, 10*time.Second)
|
|
|
|
// Verify ephemeral peer is present before disconnect.
|
|
_, found := regular.PeerByName("eph-cleanup-ephemeral")
|
|
require.True(t, found, "ephemeral peer should be visible before disconnect")
|
|
|
|
// Ensure the ephemeral node's long-poll session is fully
|
|
// established on the server before disconnecting. Without
|
|
// this, the [TestClient.Disconnect] may cancel a [controlclient.Direct.PollNetMap] that hasn't
|
|
// yet reached serveLongPoll, so no grace period or ephemeral
|
|
// GC would ever be scheduled.
|
|
ephemeral.WaitForPeers(t, 1, 10*time.Second)
|
|
|
|
// Disconnect the ephemeral node.
|
|
ephemeral.Disconnect(t)
|
|
|
|
// After the grace period (10s) + ephemeral timeout (3s),
|
|
// the ephemeral node should be deleted from the server and
|
|
// disappear from the regular node's peer list entirely.
|
|
// Unlike non-ephemeral nodes which go offline but stay in
|
|
// the peer list, ephemeral nodes should be garbage collected.
|
|
regular.WaitForCondition(t, "ephemeral peer removed from peer list",
|
|
60*time.Second,
|
|
func(nm *netmap.NetworkMap) bool {
|
|
for _, p := range nm.Peers {
|
|
hi := p.Hostinfo()
|
|
if hi.Valid() && hi.Hostname() == "eph-cleanup-ephemeral" {
|
|
return false // still present
|
|
}
|
|
}
|
|
|
|
return true // gone
|
|
})
|
|
})
|
|
|
|
t.Run("ephemeral_and_regular_mixed", func(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
srv := servertest.NewServer(t,
|
|
servertest.WithEphemeralTimeout(5*time.Second))
|
|
user := srv.CreateUser(t, "mix-user")
|
|
|
|
r1 := servertest.NewClient(t, srv, "mix-regular-1",
|
|
servertest.WithUser(user))
|
|
r2 := servertest.NewClient(t, srv, "mix-regular-2",
|
|
servertest.WithUser(user))
|
|
e1 := servertest.NewClient(t, srv, "mix-eph-1",
|
|
servertest.WithUser(user), servertest.WithEphemeral())
|
|
|
|
// All three should see each other.
|
|
r1.WaitForPeers(t, 2, 15*time.Second)
|
|
r2.WaitForPeers(t, 2, 15*time.Second)
|
|
e1.WaitForPeers(t, 2, 15*time.Second)
|
|
|
|
servertest.AssertMeshComplete(t,
|
|
[]*servertest.TestClient{r1, r2, e1})
|
|
})
|
|
|
|
t.Run("ephemeral_reconnect_prevents_cleanup", func(t *testing.T) {
|
|
t.Parallel()
|
|
|
|
srv := servertest.NewServer(t,
|
|
servertest.WithEphemeralTimeout(5*time.Second))
|
|
user := srv.CreateUser(t, "eph-recon-user")
|
|
|
|
regular := servertest.NewClient(t, srv, "eph-recon-regular",
|
|
servertest.WithUser(user))
|
|
ephemeral := servertest.NewClient(t, srv, "eph-recon-ephemeral",
|
|
servertest.WithUser(user), servertest.WithEphemeral())
|
|
|
|
regular.WaitForPeers(t, 1, 10*time.Second)
|
|
|
|
// Ensure the ephemeral node's long-poll is established.
|
|
ephemeral.WaitForPeers(t, 1, 10*time.Second)
|
|
|
|
// Disconnect and quickly reconnect.
|
|
ephemeral.Disconnect(t)
|
|
ephemeral.Reconnect(t)
|
|
|
|
// After reconnecting, the ephemeral node should still be visible.
|
|
regular.WaitForPeers(t, 1, 15*time.Second)
|
|
|
|
_, found := regular.PeerByName("eph-recon-ephemeral")
|
|
assert.True(t, found,
|
|
"ephemeral node should still be visible after quick reconnect")
|
|
})
|
|
}
|