Eliminate get_committed_snapshots from join path.

tests/code_update.py

@@ -371,7 +371,7 @@ def test_tcb_version_tables(network, args):
     thrown_exception = None
     try:
         new_node = network.create_node()
-        network.join_node(new_node, args.package, args, timeout=3)
+        network.join_node(new_node, args.package, args, timeout=3, from_snapshot=False)
         network.trust_node(new_node, args)
     except TimeoutError as e:
         thrown_exception = e
@@ -392,7 +392,7 @@ def test_tcb_version_tables(network, args):
 
     LOG.info("Checking new nodes are allowed to join using expanded api")
     new_node = network.create_node()
-    network.join_node(new_node, args.package, args, timeout=3)
+    network.join_node(new_node, args.package, args, timeout=3, from_snapshot=False)
     network.trust_node(new_node, args)
 
     LOG.info("Change the current cpuid's TCB version using the new API")
@@ -413,7 +413,7 @@ def test_tcb_version_tables(network, args):
 
     LOG.info("Checking new nodes are allowed to join using hexstring api")
     new_node = network.create_node()
-    network.join_node(new_node, args.package, args, timeout=3)
+    network.join_node(new_node, args.package, args, timeout=3, from_snapshot=False)
     network.trust_node(new_node, args)
 
 
@@ -433,6 +433,7 @@ def test_add_node_without_security_policy(network, args):
             args,
             timeout=3,
             snp_uvm_security_context_dir=snp_dir if security_context_dir else None,
+            from_snapshot=False,
         )
         network.trust_node(new_node, args)
         return network
@@ -459,7 +460,7 @@ def test_add_node_with_stubbed_security_policy(network, args):
 
     # If we don't throw an exception, joining was successful
     new_node = network.create_node()
-    network.join_node(new_node, args.package, args, timeout=3)
+    network.join_node(new_node, args.package, args, timeout=3, from_snapshot=False)
     network.trust_node(new_node, args)
 
     # Revert to original state
@@ -494,6 +495,7 @@ def test_start_node_with_mismatched_host_data(network, args):
                 args,
                 timeout=3,
                 snp_uvm_security_context_dir=snp_dir if security_context_dir else None,
+                from_snapshot=False,
             )
     except (TimeoutError, RuntimeError):
         LOG.info("As expected, node with invalid security policy failed to startup")
@@ -518,7 +520,7 @@ def test_add_node_with_untrusted_measurement(network, args):
 
     new_node = network.create_node()
     try:
-        network.join_node(new_node, args.package, args, timeout=3)
+        network.join_node(new_node, args.package, args, timeout=3, from_snapshot=False)
     except infra.network.MeasurementNotFound:
         LOG.info("As expected, node with untrusted measurement failed to join")
     else:
@@ -586,7 +588,7 @@ def assert_node_join_fails(network, args):
     """Create a node and assert that joining the network raises HostDataNotFound."""
     new_node = network.create_node()
     try:
-        network.join_node(new_node, args.package, args, timeout=3)
+        network.join_node(new_node, args.package, args, timeout=3, from_snapshot=False)
     except infra.network.HostDataNotFound as e:
         LOG.info(f"As expected, node join failed: {e.error_line}")
         assert (
@@ -724,7 +726,9 @@ def test_add_node_via_code_policy(network, args):
         ),
     )
     new_node = network.create_node()
-    network.join_node(new_node, joiner_args.package, joiner_args, timeout=3)
+    network.join_node(
+        new_node, joiner_args.package, joiner_args, timeout=3, from_snapshot=False
+    )
     network.trust_node(new_node, joiner_args)
 
     # Cleanup: restore host data and remove code update policy.
@@ -754,7 +758,7 @@ def test_add_node_with_untrusted_host_data(network, args):
 
     new_node = network.create_node()
     try:
-        network.join_node(new_node, args.package, args, timeout=3)
+        network.join_node(new_node, args.package, args, timeout=3, from_snapshot=False)
     except infra.network.HostDataNotFound:
         LOG.info("As expected, node with untrusted host data failed to join")
     else:
@@ -788,6 +792,7 @@ def test_add_node_with_no_uvm_endorsements(network, args):
                 args,
                 timeout=3,
                 snp_uvm_security_context_dir=snp_dir if security_context_dir else None,
+                from_snapshot=False,
             )
         except infra.network.MeasurementNotFound:
             LOG.info("As expected, node with no UVM endorsements failed to join")
@@ -810,6 +815,7 @@ def test_add_node_with_no_uvm_endorsements(network, args):
             args,
             timeout=3,
             snp_uvm_security_context_dir=snp_dir if security_context_dir else None,
+            from_snapshot=False,
         )
         new_node.stop()
 
@@ -840,6 +846,7 @@ def test_add_node_with_different_package(network, args):
             replacement_package,
             args,
             timeout=3,
+            from_snapshot=False,
         )
 
     except (infra.network.MeasurementNotFound, infra.network.HostDataNotFound) as err:
@@ -994,7 +1001,7 @@ def format_expected_host_data(entries):
     LOG.info("Start fresh nodes running new code")
     for _ in range(0, len(old_nodes)):
         new_node = network.create_node()
-        network.join_node(new_node, replacement_package, args)
+        network.join_node(new_node, replacement_package, args, from_snapshot=False)
         network.trust_node(new_node, args)
 
     LOG.info("Retire original nodes running old code")
@@ -1075,7 +1082,7 @@ def test_add_node_with_no_uvm_endorsements_in_kv(network, args):
 
     try:
         new_node = network.create_node()
-        network.join_node(new_node, args.package, args, timeout=3)
+        network.join_node(new_node, args.package, args, timeout=3, from_snapshot=False)
     except infra.network.UVMEndorsementsNotAuthorised:
         LOG.info("As expected, node with no UVM endorsements failed to join")
     else:

tests/e2e_common_endpoints.py

@@ -33,7 +33,7 @@ def test_primary(network, args):
     host_spec.with_args(args)
 
     new_backup = network.create_node(host_spec)
-    network.join_node(new_backup, args.package, args)
+    network.join_node(new_backup, args.package, args, from_snapshot=False)
     network.trust_node(new_backup, args)
 
     primary_interfaces = primary.host.rpc_interfaces
@@ -149,7 +149,7 @@ def test_network_node_info(network, args):
     host_spec.with_args(args)
 
     new_node = network.create_node(host_spec)
-    network.join_node(new_node, args.package, args)
+    network.join_node(new_node, args.package, args, from_snapshot=False)
 
     with new_node.client(interface_name=operator_rpc_interface) as c:
         r = c.get("/node/network/nodes/self", allow_redirects=False)

tests/e2e_logging.py

@@ -1543,7 +1543,7 @@ def test_long_lived_forwarding(network, args):
 
     new_node_args = copy.deepcopy(args)
     new_node_args.node_to_node_message_limit = message_limit
-    network.join_node(new_node, args.package, new_node_args)
+    network.join_node(new_node, args.package, new_node_args, from_snapshot=False)
     network.trust_node(new_node, new_node_args)
 
     # Send many messages to new node over long-lived connections,

tests/e2e_operations.py

@@ -689,6 +689,7 @@ def test_empty_snapshot(network, args):
                 snapshots_dir=snapshots_dir,
                 # Don't try to fetch a snapshot, look at the local files
                 fetch_recent_snapshot=False,
+                from_snapshot=True,
             )
             new_node.stop()
 
@@ -728,6 +729,7 @@ def test_nulled_snapshot(network, args):
                 snapshots_dir=snapshots_dir,
                 # Don't try to fetch a snapshot, look at the local files
                 fetch_recent_snapshot=False,
+                from_snapshot=True,
             )
         except Exception as e:
             failed = True
@@ -783,6 +785,7 @@ def test_corrupt_snapshot_handling(network, args):
             args.package,
             args,
             snapshots_dir=writable_dir,
+            from_snapshot=True,
             fetch_recent_snapshot=False,
         )
 
@@ -875,6 +878,7 @@ def test_corrupt_snapshot_handling(network, args):
             args.package,
             args,
             snapshots_dir=restricted_dir,
+            from_snapshot=True,
             fetch_recent_snapshot=False,
         )
 
@@ -1424,10 +1428,15 @@ def test_ledger_chunk_redirect_gap(network, args):
         commit_seqno = TxID.from_str(r["transaction_id"]).seqno
 
     new_node = network.create_node()
+    # force primary to generate a new snapshot after commit idx
+    snapshots = network.get_committed_snapshots()
     network.join_node(
         new_node,
         args.package,
         args,
+        # Tmp fix until the primary expresses an opinion
+        snapshots_dir=snapshots,
+        from_snapshot=True,
         # Fetch recent snapshot to speed up joining
         fetch_recent_snapshot=True,
     )
@@ -1591,7 +1600,7 @@ def run_tls_san_checks(const_args):
         )
         new_node = network.create_node(host_spec)
         args.subject_alt_names = [f"dNSName:{dummy_san}"]
-        network.join_node(new_node, args.package, args)
+        network.join_node(new_node, args.package, args, from_snapshot=False)
         sans = infra.crypto.get_san_from_pem_cert(new_node.get_tls_certificate_pem())
         assert len(sans) == 1, "Expected exactly one SAN"
         assert sans[0].value == dummy_san
@@ -1606,7 +1615,7 @@ def run_tls_san_checks(const_args):
             dummy_public_rpc_hosts.add(ipaddress.ip_address(dummy_public_rpc_host))
 
         new_node = network.create_node(host_spec)
-        network.join_node(new_node, args.package, args)
+        network.join_node(new_node, args.package, args, from_snapshot=False)
         # Cannot trust the node here as client cannot authenticate dummy public IP in cert
         with open(
             os.path.join(network.common_dir, f"{new_node.local_node_id}.pem"),
@@ -1941,7 +1950,7 @@ def replace_nodes(network, old_nodes, new_package):
         new_nodes = []
         for _ in range(len(old_nodes)):
             new_node = network.create_node()
-            network.join_node(new_node, new_package, nargs)
+            network.join_node(new_node, new_package, nargs, from_snapshot=False)
             network.trust_node(new_node, nargs)
             new_nodes.append(new_node)
 
@@ -1999,7 +2008,7 @@ def replace_nodes(network, old_nodes, new_package):
         # Verify a Dual joiner can still join (allow_dual_signing_joinee=true)
         LOG.info("Verifying Dual joiner can still join COSE-only (allow dual) network")
         dual_joiner = network.create_node()
-        network.join_node(dual_joiner, nargs.package, nargs)
+        network.join_node(dual_joiner, nargs.package, nargs, from_snapshot=False)
         network.trust_node(dual_joiner, nargs)
         LOG.success("Dual joiner successfully joined COSE-only network")
 
@@ -2039,7 +2048,9 @@ def replace_nodes(network, old_nodes, new_package):
         LOG.info("Verifying Dual joiner is rejected by COSE-only-no-dual network")
         rejected_joiner = network.create_node()
         try:
-            network.join_node(rejected_joiner, nargs.package, nargs, timeout=10)
+            network.join_node(
+                rejected_joiner, nargs.package, nargs, timeout=10, from_snapshot=False
+            )
             network.trust_node(rejected_joiner, nargs)
             assert False, "Dual joiner should have been rejected"
         except Exception as e:
@@ -2185,11 +2196,14 @@ def try_historical_fetch(node, timeout=1):
         # Create a temporary directory to manually construct a ledger in
         with tempfile.TemporaryDirectory() as temp_dir:
             new_node = network.create_node()
+
+            snapshots_dir = network.get_committed_snapshots()
             network.join_node(
                 new_node,
                 nargs.package,
                 nargs,
                 from_snapshot=True,
+                snapshots_dir=snapshots_dir,
                 copy_ledger=False,
                 common_read_only_ledger_dir=temp_dir,  # New node will try to read from temp directory
             )
@@ -3004,7 +3018,9 @@ def run_error_message_on_failure_to_read_aci_sec_context(args):
         args_copy.snp_endorsements_file = "/a/fake/path"
         failed = False
         try:
-            network.join_node(new_node, args.package, args_copy, timeout=20)
+            network.join_node(
+                new_node, args.package, args_copy, timeout=20, from_snapshot=False
+            )
         except infra.network.CollateralFetchTimeout:
             LOG.info(
                 "Node with invalid quote endorsement servers could not join as expected"
@@ -3175,6 +3191,7 @@ def test_backup_snapshot_fetch_max_size(network, args):
         args,
         target_node=primary,
         timeout=5,
+        from_snapshot=False,
         backup_snapshot_fetch_enabled=True,
         backup_snapshot_fetch_max_attempts=1,
         backup_snapshot_fetch_max_size="1KB",

tests/governance.py

@@ -133,7 +133,7 @@ def test_no_quote(network, args):
             }
         )
     )
-    network.join_node(untrusted_node, args.package, args)
+    network.join_node(untrusted_node, args.package, args, from_snapshot=False)
     with untrusted_node.client(
         ca=os.path.join(
             untrusted_node.common_dir, f"{untrusted_node.local_node_id}.pem"
@@ -175,7 +175,7 @@ def get_nodes():
             )
 
             # NB: This new node joins but is never trusted
-            network.join_node(untrusted_node, args.package, args)
+            network.join_node(untrusted_node, args.package, args, from_snapshot=False)
 
             nodes = get_nodes()
             assert untrusted_node.node_id in nodes, nodes

tests/infra/basicperf.py

@@ -240,6 +240,7 @@ def replace_primary(network, host, old_primary, snapshots_dir, statistics):
         timeout=10,
         copy_ledger=False,
         snapshots_dir=snapshots_dir,
+        from_snapshot=snapshots_dir is not None,
         follow_redirect=False,
     )
     LOG.info(f"Shut down primary: {old_primary.local_node_id}")

tests/infra/network.py

@@ -350,7 +350,7 @@ def _setup_node(
         ledger_dir=None,
         copy_ledger=True,
         read_only_ledger_dirs=None,
-        from_snapshot=True,
+        from_snapshot=False,
         snapshots_dir=None,
         **kwargs,
     ):
@@ -368,20 +368,16 @@ def _setup_node(
 
         # Note: Copy snapshot before ledger as retrieving the latest snapshot may require
         # to produce more ledger entries
+        # Note: from_snapshot is not true in the start_and_open case nor start_in_recovery
         if from_snapshot:
             # Only retrieve snapshot from primary if the snapshot directory is not specified
             if snapshots_dir is None:
-                primary, _ = self.find_primary(
-                    timeout=args.ledger_recovery_timeout if recovery else 10
-                )
-                read_only_snapshots_dir = self.get_committed_snapshots(primary)
-            if os.listdir(snapshots_dir) or os.listdir(read_only_snapshots_dir):
-                LOG.info(
-                    f"Joining from snapshot directories: {snapshots_dir},{read_only_snapshots_dir}"
-                )
+                assert False, "snapshot_dir must be provided when from_snapshot is True"
+            if snapshots_dir and os.listdir(snapshots_dir):
+                LOG.info(f"Joining from snapshot directories: {snapshots_dir}")
             else:
                 LOG.warning(
-                    f"Attempting to join from snapshot but {snapshots_dir},{read_only_snapshots_dir} are empty: defaulting to complete replay of transaction history"
+                    f"Attempting to join from snapshot but {snapshots_dir} is empty: defaulting to complete replay of transaction history"
                 )
         else:
             LOG.info(
@@ -424,7 +420,7 @@ def _add_node(
         ledger_dir=None,
         copy_ledger=True,
         read_only_ledger_dirs=None,
-        from_snapshot=True,
+        from_snapshot=False,
         snapshots_dir=None,
         **kwargs,
     ):
@@ -1603,7 +1599,7 @@ def resize(self, target_count, args):
         LOG.info(f"Resizing network from {initial_node_count} to {target_count} nodes")
         while node_count < target_count:
             new_node = self.create_node()
-            self.join_node(new_node, args.package, args)
+            self.join_node(new_node, args.package, args, from_snapshot=False)
             self.trust_node(new_node, args)
             node_count += 1
         while node_count > target_count:
@@ -1870,12 +1866,14 @@ def wait_for_primary_unanimity(
 
     def get_committed_snapshots(
         self,
-        node,
+        node=None,
         target_seqno=None,
         force_txs=True,
         wait_for_target_seqno=False,
         timeout=20,
     ):
+        if node is None:
+            node, _ = self.find_primary()
         # Wait for the snapshot including target_seqno to be committed before
         # copying snapshot directory. Do not issue transactions if force_txs is False
         # and expect snapshot to have already been created.

tests/js-custom-authorization/custom_authorization.py

@@ -1092,6 +1092,7 @@ def test_metrics_logging(network, args):
         new_node,
         args.package,
         args,
+        from_snapshot=False,
     )
     network.trust_node(new_node, args)
 

tests/limits.py

@@ -25,7 +25,7 @@ def test_forward_larger_than_default_requests(network, args):
             }
         )
     )
-    network.join_node(new_node, args.package, args)
+    network.join_node(new_node, args.package, args, from_snapshot=False)
     network.trust_node(new_node, args)
 
     primary, _ = network.find_primary()