fix(upgrade_test): ignore topology change coordinator errors during u…

…pgrades

The error messages reported in scylladb#20754 and scylladb#20950 can be
ignored till a proper fix is merged, because it's mostly a bad UX.

sdcm/sct_events/group_common_events.py

@@ -97,6 +97,36 @@ def ignore_upgrade_schema_errors():
             line="cql_server - exception while processing connection: seastar::nested_exception "
                  "(seastar::nested_exception)",
         ))
+        if SkipPerIssues(
+            issues=[
+                "https://github.com/scylladb/scylladb/issues/20754",
+                "https://github.com/scylladb/scylladb/issues/20950",
+            ],
+            params=TestConfig().tester_obj().params,
+        ):
+            # @piodul:
+            #
+            #   The upgrade-to-view-build-status-on-raft procedure runs right after the VIEW_BUILD_STATUS_ON_GROUP0
+            #   feature is enabled.  Enabling a cluster feature on raft requires all nodes to be alive.  Most likely
+            #   the node being restarted wasn't yet seen as such, but the upgrade procedure started anyway.
+            #
+            #   The error is not critical.  The topology coordinator node will retry the operation in short intervals
+            #   until it succeeds.  The operation shouldn't have any harmful side effects if it fails, so it's mostly
+            #   bad UX because we can avoid the busywork and error messages by appropriately delaying the moment when
+            #   the operation is executed.
+            #
+            #   Therefore, it is OK to ignore this particular error until a proper fix is merged.
+            stack.enter_context(DbEventsFilter(
+                db_event=DatabaseLogEvent.DATABASE_ERROR,
+                line="raft_topology - topology change coordinator fiber got error exceptions::unavailable_exception"
+                     " (Cannot achieve consistency level for cl ALL.",
+            ))
+            stack.enter_context(DbEventsFilter(
+                db_event=DatabaseLogEvent.RUNTIME_ERROR,
+                line="raft_topology - topology change coordinator fiber got error std::runtime_error"
+                     " (raft topology: exec_global_command(barrier) failed with seastar::rpc::closed_error"
+                     " (connection is closed))",
+            ))
         yield