hostname validation in InstanceSerializer (#12979)

* initial commit of hostname validation to InstanceSerializer

Co-authored-by: Cesar Francisco San Nicolas Martinez <cesarfsannicolasmartinez@gmail.com>

awx/api/serializers.py

@@ -29,6 +29,7 @@ from django.utils.translation import gettext_lazy as _
 from django.utils.encoding import force_str
 from django.utils.text import capfirst
 from django.utils.timezone import now
+from django.core.validators import RegexValidator, MaxLengthValidator
 
 # Django REST Framework
 from rest_framework.exceptions import ValidationError, PermissionDenied
@@ -120,6 +121,9 @@ from awx.main.validators import vars_validate_or_raise
 from awx.api.versioning import reverse
 from awx.api.fields import BooleanNullField, CharNullField, ChoiceNullField, VerbatimField, DeprecatedCredentialField
 
+# AWX Utils
+from awx.api.validators import HostnameRegexValidator
+
 logger = logging.getLogger('awx.api.serializers')
 
 # Fields that should be summarized regardless of object type.
@@ -4921,6 +4925,18 @@ class InstanceSerializer(BaseSerializer):
         extra_kwargs = {
             'node_type': {'initial': Instance.Types.EXECUTION, 'default': Instance.Types.EXECUTION},
             'node_state': {'initial': Instance.States.INSTALLED, 'default': Instance.States.INSTALLED},
+            'hostname': {
+                'validators': [
+                    MaxLengthValidator(limit_value=255),
+                    RegexValidator(
+                        regex='^localhost$|^127(?:\.[0-9]+){0,2}\.[0-9]+$|^(?:0*\:)*?:?0*1$',
+                        flags=re.IGNORECASE,
+                        inverse_match=True,
+                        message="hostname cannot be localhost or 127.0.0.1",
+                    ),
+                    HostnameRegexValidator(),
+                ],
+            },
         }
 
     def get_related(self, obj):
@@ -4991,6 +5007,10 @@ class InstanceSerializer(BaseSerializer):
         return value
 
     def validate_hostname(self, value):
+        """
+        - Hostname cannot be "localhost" - but can be something like localhost.domain
+        - Cannot change the hostname of an-already instantiated & initialized Instance object
+        """
         if self.instance and self.instance.hostname != value:
             raise serializers.ValidationError("Cannot change hostname.")
 

awx/api/validators.py

@@ -0,0 +1,55 @@
+import re
+
+from django.core.validators import RegexValidator, validate_ipv46_address
+from django.core.exceptions import ValidationError
+
+
+class HostnameRegexValidator(RegexValidator):
+    """
+    Fully validates a domain name that is compliant with norms in Linux/RHEL
+        - Cannot start with a hyphen
+        - Cannot begin with, or end with a "."
+        - Cannot contain any whitespaces
+        - Entire hostname is max 255 chars (including dots)
+        - Each domain/label is between 1 and 63 characters, except top level domain, which must be at least 2 characters
+        - Supports ipv4, ipv6, simple hostnames and FQDNs
+        - Follows RFC 9210 (modern RFC 1123, 1178) requirements
+
+    Accepts an IP Address or Hostname as the argument
+    """
+
+    regex = '^[a-z0-9][-a-z0-9]*$|^([a-z0-9][-a-z0-9]{0,62}[.])*[a-z0-9][-a-z0-9]{1,62}$'
+    flags = re.IGNORECASE
+
+    def __call__(self, value):
+        regex_matches, err = self.__validate(value)
+        invalid_input = regex_matches if self.inverse_match else not regex_matches
+        if invalid_input:
+            if err is None:
+                err = ValidationError(self.message, code=self.code, params={"value": value})
+            raise err
+
+    def __str__(self):
+        return f"regex={self.regex}, message={self.message}, code={self.code}, inverse_match={self.inverse_match}, flags={self.flags}"
+
+    def __validate(self, value):
+
+        if ' ' in value:
+            return False, ValidationError("whitespaces in hostnames are illegal")
+
+        """
+        If we have an IP address, try and validate it.
+        """
+        try:
+            validate_ipv46_address(value)
+            return True, None
+        except ValidationError:
+            pass
+
+        """
+        By this point in the code, we probably have a simple hostname, FQDN or a strange hostname like "192.localhost.domain.101"
+        """
+        if not self.regex.match(value):
+            return False, ValidationError(f"illegal characters detected in hostname={value}. Please verify.")
+
+        return True, None