Fix Gemma 4 vLLM startup and improve partial-provisioning UX

- provisioning.rb: auto-install pytest before pre_start_cmd to fix missing
dependency in nightly vLLM images that crashes EngineCore on cupy import
- cli.rb: relax active_config_loaders/vm1_alive? to match by vm_id+public_ip
instead of requiring status==ACTIVE, so watch/status/test work on VMs
that were interrupted mid-provisioning
- vm_lifecycle.rb: persist API status/vm_state back to state file during
status calls; add 'Provisioning: incomplete' notice when provisioned_at
is missing
- provisioning_orchestrator.rb: add wait_for_ssh polling before
ensure_trusted_host to fix race between SSH daemon readiness and keyscan

1 files changed, 7 insertions, 0 deletions

diff --git a/lib/hyperstack/provisioning_orchestrator.rb b/lib/hyperstack/provisioning_orchestrator.rb
index 7dbb10e..b9f9e3d 100644
--- a/lib/hyperstack/provisioning_orchestrator.rb
+++ b/lib/hyperstack/provisioning_orchestrator.rb
@@ -29,6 +29,7 @@ module HyperstackVM
       state['security_rules'] = Array(vm['security_rules']).map { |r| normalize_rule(r) }
       @state_store.save(state)
 
+      wait_for_ssh(state['public_ip'])
       @ssh_runner.ensure_trusted_host(state['public_ip'])
 
       if @config.guest_bootstrap_enabled? && state['bootstrapped_at'].nil?
@@ -78,6 +79,12 @@ module HyperstackVM
       state
     end
 
+    def wait_for_ssh(host)
+      with_polling("SSH on #{host}:#{@config.ssh_port} to become reachable", timeout: 300) do
+        @ssh_runner.tcp_open?(host, @config.ssh_port)
+      end
+    end
+
     def wait_for_ready(vm_id)
       with_polling("VM #{vm_id} to become ready for firewall updates") do
         vm = @client.get_vm(vm_id)