agentHost: improve update pickup and fix startup-failure wedge (#313541)

- Shorten background update check interval from 24h to 6h so newly
  published server versions are picked up sooner after auto-shutdown.
- Don't store the child into self.running when it exits before signaling
  ready. Previously a failed startup left a dead child in running
  forever, wedging ensure_server() so it could never restart the server.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

cli/src/tunnels/agent_host.rs

@@ -33,7 +33,7 @@ use crate::util::sync::{new_barrier, Barrier, BarrierOpener};
 use super::paths::{get_server_folder_name, SERVER_FOLDER_NAME};
 
 /// How often to check for server updates.
-pub const UPDATE_CHECK_INTERVAL: Duration = Duration::from_secs(24 * 60 * 60);
+pub const UPDATE_CHECK_INTERVAL: Duration = Duration::from_secs(6 * 60 * 60);
 /// How often to re-check whether the server has exited when an update is pending.
 pub const UPDATE_POLL_INTERVAL: Duration = Duration::from_secs(10 * 60);
 /// How long to wait for the server to signal readiness.
@@ -243,7 +243,10 @@ impl AgentHostManager {
 					if let Some(o) = opener.take() {
 						o.open(Err(format!("Server exited before ready: {e:?}")));
 					}
-					break;
+					// Child has already exited; don't store it in `running`,
+					// otherwise the manager would be wedged with a dead child
+					// forever and ensure_server() would never restart.
+					return;
 				}
 			}
 
@@ -261,10 +264,6 @@ impl AgentHostManager {
 			});
 		}
 
-		if !ready {
-			return;
-		}
-
 		info!(self.log, "[{}]: Server ready", commit_prefix);
 
 		// Continue reading output until the process exits