diff --git a/src/opnsense/scripts/filter/lib/alias.py b/src/opnsense/scripts/filter/lib/alias.py
index 9e8341be4..b2b0596a0 100644
--- a/src/opnsense/scripts/filter/lib/alias.py
+++ b/src/opnsense/scripts/filter/lib/alias.py
@@ -147,8 +147,12 @@ class Alias(object):
                     if raw_address and not raw_address.startswith('//'):
                         for address in self._parse_address(raw_address):
                             yield address
+            else:
+                syslog.syslog(syslog.LOG_ERR, 'error fetching alias url %s [http_code:%s]' % (url, req.status_code))
+                raise IOError('error fetching alias url %s' % (url))
         except:
             syslog.syslog(syslog.LOG_ERR, 'error fetching alias url %s' % (url))
+            raise IOError('error fetching alias url %s' % (url))
 
     def _fetch_geo(self, geoitem):
         """ fetch geoip addresses, if not downloaded or outdated force an update
@@ -218,20 +222,30 @@ class Alias(object):
         """
         if not self._resolve_content:
             if self.expired() or self.changed() or force:
-                with open(self._filename_alias_content, 'w') as f_out:
-                    for item in self.items():
-                        address_parser = self.get_parser()
-                        if address_parser:
-                            for address in address_parser(item):
-                                if address not in self._resolve_content:
-                                    # flush new alias content (without dependencies) to disk, so progress can easliy
-                                    # be followed, large lists of domain names can take quite some resolve time.
-                                    f_out.write('%s\n' % address)
-                                    f_out.flush()
-                                    # preserve addresses
-                                    self._resolve_content.append(address)
-                    # flush md5 hash to disk
-                    open(self._filename_alias_hash, 'w').write(self.uniqueid())
+                if os.path.isfile(self._filename_alias_content):
+                    undo_content = open(self._filename_alias_content, 'r').read()
+                else:
+                    undo_content = ""
+                try:
+                    with open(self._filename_alias_content, 'w') as f_out:
+                        for item in self.items():
+                            address_parser = self.get_parser()
+                            if address_parser:
+                                for address in address_parser(item):
+                                    if address not in self._resolve_content:
+                                        # flush new alias content (without dependencies) to disk, so progress can easliy
+                                        # be followed, large lists of domain names can take quite some resolve time.
+                                        f_out.write('%s\n' % address)
+                                        f_out.flush()
+                                        # preserve addresses
+                                        self._resolve_content.append(address)
+                except IOError:
+                    # parse issue, keep data as-is, flush previous content to disk
+                    with open(self._filename_alias_content, 'w') as f_out:
+                        f_out.write(undo_content)
+                    self._resolve_content = undo_content.split("\n")
+                # flush md5 hash to disk
+                open(self._filename_alias_hash, 'w').write(self.uniqueid())
             else:
                 self._resolve_content = open(self._filename_alias_content).read().split()
         # return the addresses and networks of this alias