DataSourceAzure: Adjust timeout for polling IMDS

If the IMDS primary server is not available, falling back to the
secondary server takes about 1s. The net result is that the
expected E2E time is slightly more than 1s. This change increases
the timeout to 2s to prevent the infinite loop of timeouts.

2 files changed, 17 insertions, 8 deletions

diff --git a/cloudinit/sources/DataSourceAzure.py b/cloudinit/sources/DataSourceAzure.py
index 64165259..b7440c1d 100755
--- a/cloudinit/sources/DataSourceAzure.py
+++ b/cloudinit/sources/DataSourceAzure.py
@@ -57,7 +57,12 @@ AZURE_CHASSIS_ASSET_TAG = '7783-7084-3265-9085-8269-3286-77'
 REPROVISION_MARKER_FILE = "/var/lib/cloud/data/poll_imds"
 REPORTED_READY_MARKER_FILE = "/var/lib/cloud/data/reported_ready"
 AGENT_SEED_DIR = '/var/lib/waagent'
+
+# In the event where the IMDS primary server is not
+# available, it takes 1s to fallback to the secondary one
+IMDS_TIMEOUT_IN_SECONDS = 2
 IMDS_URL = "http://169.254.169.254/metadata/"
+
 PLATFORM_ENTROPY_SOURCE = "/sys/firmware/acpi/tables/OEM0"
 
 # List of static scripts and network config artifacts created by
@@ -582,9 +587,9 @@ class DataSourceAzure(sources.DataSource):
                         return
                     self._ephemeral_dhcp_ctx.clean_network()
                 else:
-                    return readurl(url, timeout=1, headers=headers,
-                                   exception_cb=exc_cb, infinite=True,
-                                   log_req_resp=False).contents
+                    return readurl(url, timeout=IMDS_TIMEOUT_IN_SECONDS,
+                                   headers=headers, exception_cb=exc_cb,
+                                   infinite=True, log_req_resp=False).contents
             except UrlError:
                 # Teardown our EphemeralDHCPv4 context on failure as we retry
                 self._ephemeral_dhcp_ctx.clean_network()
@@ -1291,8 +1296,8 @@ def _get_metadata_from_imds(retries):
     headers = {"Metadata": "true"}
     try:
         response = readurl(
-            url, timeout=1, headers=headers, retries=retries,
-            exception_cb=retry_on_url_exc)
+            url, timeout=IMDS_TIMEOUT_IN_SECONDS, headers=headers,
+            retries=retries, exception_cb=retry_on_url_exc)
     except Exception as e:
         LOG.debug('Ignoring IMDS instance metadata: %s', e)
         return {}
diff --git a/tests/unittests/test_datasource/test_azure.py b/tests/unittests/test_datasource/test_azure.py
index ab77c034..427ab7e7 100644
--- a/tests/unittests/test_datasource/test_azure.py
+++ b/tests/unittests/test_datasource/test_azure.py
@@ -163,7 +163,8 @@ class TestGetMetadataFromIMDS(HttprettyTestCase):
 
         m_readurl.assert_called_with(
             self.network_md_url, exception_cb=mock.ANY,
-            headers={'Metadata': 'true'}, retries=2, timeout=1)
+            headers={'Metadata': 'true'}, retries=2,
+            timeout=dsaz.IMDS_TIMEOUT_IN_SECONDS)
 
     @mock.patch('cloudinit.url_helper.time.sleep')
     @mock.patch(MOCKPATH + 'net.is_up')
@@ -1791,7 +1792,8 @@ class TestAzureDataSourcePreprovisioning(CiTestCase):
                                     headers={'Metadata': 'true',
                                              'User-Agent':
                                              'Cloud-Init/%s' % vs()
-                                             }, method='GET', timeout=1,
+                                             }, method='GET',
+                                    timeout=dsaz.IMDS_TIMEOUT_IN_SECONDS,
                                     url=full_url)])
         self.assertEqual(m_dhcp.call_count, 2)
         m_net.assert_any_call(
@@ -1828,7 +1830,9 @@ class TestAzureDataSourcePreprovisioning(CiTestCase):
                                     headers={'Metadata': 'true',
                                              'User-Agent':
                                              'Cloud-Init/%s' % vs()},
-                                    method='GET', timeout=1, url=full_url)])
+                                    method='GET',
+                                    timeout=dsaz.IMDS_TIMEOUT_IN_SECONDS,
+                                    url=full_url)])
         self.assertEqual(m_dhcp.call_count, 2)
         m_net.assert_any_call(
             broadcast='192.168.2.255', interface='eth9', ip='192.168.2.9',