Guides image + draft gating, audit retention, ADR-015 gate, backup credential redaction

plugins/backups/migrations/versions/0002_backups_lastseenat.py

@@ -27,8 +27,10 @@ depends_on = None
 
 
 def upgrade():
-    columns = {c['name'] for c in
-               sa.inspect(op.get_bind()).get_columns('backuprevisions')}
+    inspector = sa.inspect(op.get_bind())
+    if 'backuprevisions' not in inspector.get_table_names():
+        return
+    columns = {c['name'] for c in inspector.get_columns('backuprevisions')}
     if 'lastseenat' not in columns:
         op.add_column('backuprevisions',
                       sa.Column('lastseenat', sa.DateTime(), nullable=True))
@@ -37,7 +39,9 @@ def upgrade():
 
 
 def downgrade():
-    columns = {c['name'] for c in
-               sa.inspect(op.get_bind()).get_columns('backuprevisions')}
+    inspector = sa.inspect(op.get_bind())
+    if 'backuprevisions' not in inspector.get_table_names():
+        return
+    columns = {c['name'] for c in inspector.get_columns('backuprevisions')}
     if 'lastseenat' in columns:
         op.drop_column('backuprevisions', 'lastseenat')

plugins/backups/plugin.py

@@ -410,7 +410,13 @@ class BackupsPlugin(BasePlugin):
         # with_for_update serializes concurrent posts for the same chain: the
         # fleet collects on one GE-Enforce cycle, so a retry could otherwise
         # pass the hash check twice and insert the same revision.
-        sourcehostname = payload.get('sourcehostname')
+        # Normalize ONCE, and use the same value to query and to store.
+        # samesource() maps '' to IS NULL, but the insert below used to store
+        # the raw '' - so a client posting an empty hostname wrote rows that its
+        # own next post could never match. Every post then deduped against
+        # nothing, inserting forever: unbounded growth, and a history no
+        # technician can read. Whitespace does the same thing more quietly.
+        sourcehostname = (payload.get('sourcehostname') or '').strip() or None
         latest = (db.session.query(BackupRevision)
                   .filter(BackupRevision.assetid == assetid,
                           BackupRevision.backupkind == kindkey,
@@ -476,7 +482,7 @@ class BackupsPlugin(BasePlugin):
             sharepath=sharepath,
             sourcefilename=sourcefilename,
             bytesize=bytesize,
-            sourcehostname=payload.get('sourcehostname'),
+            sourcehostname=sourcehostname,
             collectedat=collectedat or datetime.utcnow(),
             lastseenat=datetime.utcnow(),
         )

plugins/backups/services/ntlars.py

@@ -32,6 +32,31 @@ import re
 
 SCHEMA = 'ntlars/1'
 
+# SECRETS ARE NOT BACKED UP.
+#
+# The DNC registry carries an FTP password in cleartext. Stored verbatim it
+# reached three places it must not: the backuprevisions table, the revision
+# panel (which renders values), and the download route (which hands back a
+# ready-to-import .reg). Anyone with backups.view could read the fleet's FTP
+# password off a web page.
+#
+# Redaction happens at PARSE, so the secret never enters the database at all.
+# Redacting only on display would leave it at rest and one query away.
+#
+# What is deliberately NOT done: dropping the value silently. This module's
+# parser refuses to discard an unparseable value precisely because a backup that
+# looks complete and restores a machine with a setting missing is the worst
+# outcome. A redacted value keeps its NAME and records itself in
+# `redacted`, so both the panel and a restore can say plainly that this one
+# field must be set by hand.
+REDACTEDMARKER = '***REDACTED***'
+SECRETNAMEPATTERN = re.compile(r'passw|pwd|secret|apikey|token', re.IGNORECASE)
+
+
+def issecretname(name):
+    """Whether a registry value name holds a credential."""
+    return bool(SECRETNAMEPATTERN.search(name or ''))
+
 REGROOT = 'HKEY_LOCAL_MACHINE'
 DNCPATH = r'SOFTWARE\GE Aircraft Engines\DNC'
 DNCPATHWOW = r'SOFTWARE\WOW6432Node\GE Aircraft Engines\DNC'
@@ -245,6 +270,20 @@ def parse(raw):
 
     dialect = 'wow6432node' if sawwow else ('ntlars' if sawplain else 'unknown')
 
+    # Strip credentials before the projection is returned - i.e. before anything
+    # persists it. Recorded by path\name so the omission is visible rather than
+    # discovered during a restore.
+    redacted = []
+    for path in keys:
+        for name, value in keys[path].items():
+            if not issecretname(name):
+                continue
+            if value.get('data') in ('', None):
+                continue   # nothing to hide
+            keys[path][name] = {'type': value.get('type', 'REG_SZ'),
+                                'data': REDACTEDMARKER}
+            redacted.append('{}\\{}'.format(path, name) if path else name)
+
     return {
         'schema': SCHEMA,
         'sourcedialect': dialect,
@@ -252,6 +291,7 @@ def parse(raw):
             {'path': path, 'values': dict(sorted(keys[path].items()))}
             for path in sorted(keys)
         ],
+        'redacted': sorted(redacted),
     }
 
 
@@ -291,6 +331,12 @@ def render(projection, dialect='ntlars', comments=None):
 
     dialect='ntlars'      omits WOW6432Node - use with the NTLARS Load... button
     dialect='wow6432node' includes it - use with `reg import` on 64-bit
+
+    Redacted credentials are rendered as a comment naming the value, NOT as the
+    marker string. Importing '***REDACTED***' as the FTP password would leave a
+    machine that looks restored and cannot connect, and the technician would be
+    debugging the network. A commented-out value fails loudly at the one moment
+    somebody is looking at the file.
     """
     if dialect not in ('ntlars', 'wow6432node'):
         raise ValueError('unknown dialect: {!r}'.format(dialect))
@@ -302,10 +348,28 @@ def render(projection, dialect='ntlars', comments=None):
     if comments:
         out.append('')
 
+    redacted = projection.get('redacted') or []
+    if redacted:
+        out.append('; ---------------------------------------------------------------')
+        out.append('; {} credential value(s) were NOT backed up and are NOT in this'
+                   .format(len(redacted)))
+        out.append('; file. Set them by hand after importing:')
+        for name in redacted:
+            out.append(';     {}'.format(name))
+        out.append('; ---------------------------------------------------------------')
+        out.append('')
+
     for key in projection.get('keys', []):
         path = key.get('path', '')
         out.append('[{}]'.format(root + ('\\' + path if path else '')))
         for name, entry in (key.get('values') or {}).items():
+            # A redacted credential is commented out, never emitted as the
+            # marker. Importing the marker as a password produces a machine that
+            # looks restored and cannot log in.
+            if (entry or {}).get('data') == REDACTEDMARKER:
+                out.append('; "{}" was not backed up - set it by hand'.format(
+                    _escape(name)))
+                continue
             out.append(_rendervalue(name, entry))
         out.append('')