Azure: Ensure platform random_seed is always serializable as JSON.

The Azure platform surfaces random bytes into /sys via Hyper-V.
Python 2.7 json.dump() raises an exception if asked to convert
a str with non-character content, and python 3.0 json.dump()
won't serialize a "bytes" value. As a result, c-i instance
data is often not written by Azure, making reboots slower (c-i
has to repeat work).

The random data is base64-encoded and then decoded into a string
(str or unicode depending on the version of Python in use). The
base64 string has just as many bits of entropy, so we're not
throwing away useful "information", but we can be certain
json.dump() will correctly serialize the bits.

1 files changed, 19 insertions, 5 deletions

diff --git a/cloudinit/sources/DataSourceAzure.py b/cloudinit/sources/DataSourceAzure.py
index eccbee5a..b4e3f061 100644
--- a/cloudinit/sources/DataSourceAzure.py
+++ b/cloudinit/sources/DataSourceAzure.py
@@ -54,6 +54,7 @@ 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'
 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
 # stock ubuntu suported images.
@@ -195,6 +196,8 @@ if util.is_FreeBSD():
         RESOURCE_DISK_PATH = "/dev/" + res_disk
     else:
         LOG.debug("resource disk is None")
+    # TODO Find where platform entropy data is surfaced
+    PLATFORM_ENTROPY_SOURCE = None
 
 BUILTIN_DS_CONFIG = {
     'agent_command': AGENT_START_BUILTIN,
@@ -1100,16 +1103,27 @@ def _check_freebsd_cdrom(cdrom_dev):
     return False
 
 
-def _get_random_seed():
+def _get_random_seed(source=PLATFORM_ENTROPY_SOURCE):
     """Return content random seed file if available, otherwise,
        return None."""
     # azure / hyper-v provides random data here
-    # TODO. find the seed on FreeBSD platform
     # now update ds_cfg to reflect contents pass in config
-    if util.is_FreeBSD():
+    if source is None:
         return None
-    return util.load_file("/sys/firmware/acpi/tables/OEM0",
-                          quiet=True, decode=False)
+    seed = util.load_file(source, quiet=True, decode=False)
+
+    # The seed generally contains non-Unicode characters. load_file puts
+    # them into a str (in python 2) or bytes (in python 3). In python 2,
+    # bad octets in a str cause util.json_dumps() to throw an exception. In
+    # python 3, bytes is a non-serializable type, and the handler load_file
+    # uses applies b64 encoding *again* to handle it. The simplest solution
+    # is to just b64encode the data and then decode it to a serializable
+    # string. Same number of bits of entropy, just with 25% more zeroes.
+    # There's no need to undo this base64-encoding when the random seed is
+    # actually used in cc_seed_random.py.
+    seed = base64.b64encode(seed).decode()
+
+    return seed
 
 
 def list_possible_azure_ds_devs():