Manage the firewall configuration on the network device managed through NAPALM. The firewall configuration is generated by Capirca.
New in version 2017.7.0.
Mircea Ulinic <ping@mirceaulinic.net>
new
capirca, napalm
unix
To install Capirca, execute: pip install capirca
.
To be able to load configuration on network devices,
it requires NAPALM library to be installed: pip install napalm
.
Please check Installation for complete details.
salt.states.netacl.
filter
(name, filter_name, filter_options=None, terms=None, prepend=True, pillar_key='acl', pillarenv=None, saltenv=None, merge_pillar=False, only_lower_merge=False, revision_id=None, revision_no=None, revision_date=True, revision_date_format='%Y/%m/%d', test=False, commit=True, debug=False)¶Generate and load the configuration of a policy filter.
The name of the policy filter.
Additional filter options. These options are platform-specific. See the complete list of options.
Dictionary of terms for this policy filter.
If not specified or empty, will try to load the configuration from the pillar,
unless merge_pillar
is set as False
.
True
When merge_pillar
is set as True
, the final list of terms generated by merging
the terms from terms
with those defined in the pillar (if any): new terms are prepended
at the beginning, while existing ones will preserve the position. To add the new terms
at the end of the list, set this argument to False
.
acl
The key in the pillar containing the default attributes values. Default: acl
.
Query the master to generate fresh pillar data on the fly, specifically from the requested pillar environment.
Included only for compatibility with
pillarenv_from_saltenv
, and is otherwise ignored.
False
Merge terms
with the corresponding value from the pillar. Default: False
.
Note
By default this state does not merge, to avoid any unexpected behaviours.
The merge logic depends on the prepend
argument.
The terms specified through the terms
argument have higher priority
than the pillar.
False
Specify if it should merge only the terms fields. Otherwise it will try
to merge also filters fields. Default: False
.
This option requires merge_pillar
, otherwise it is ignored.
Add a comment in the filter config having the description for the changes applied.
The revision count.
True
Boolean flag: display the date when the filter configuration was generated. Default: True
.
%Y/%m/%d
The date format to be used when generating the perforce data. Default: %Y/%m/%d
(<year>/<month>/<day>).
False
Dry run? If set as True
, will apply the config, discard and return the changes.
Default: False
and will commit the changes on the device.
True
Commit? Default: True
.
False
Debug mode. Will insert a new key under the output dictionary,
as loaded_config
containing the raw configuration loaded on the device.
CLI Example:
salt 'edge01.flw01' state.sls router.acl test=True
Output Example:
edge01.flw01:
----------
ID: my-filter
Function: netacl.filter
Result: None
Comment: Testing mode: Configuration discarded.
Started: 12:24:40.598232
Duration: 2437.139 ms
Changes:
----------
diff:
---
+++
@@ -1228,9 +1228,24 @@
!
+ipv4 access-list my-filter
+ 10 remark $Id: my-filter_state $
+ 20 remark $Revision: 5 $
+ 30 remark my-other-term
+ 40 permit tcp any range 5678 5680 any
+!
+!
loaded:
! $Id: my-filter_state $
! $Revision: 5 $
no ipv6 access-list my-filter
ipv6 access-list my-filter
remark $Id: my-filter_state $
remark $Revision: 5 $
remark my-other-term
permit tcp any range 5678 5680 any
exit
Summary for edge01.flw01
------------
Succeeded: 1 (unchanged=1, changed=1)
Failed: 0
------------
Total states run: 1
Total run time: 2.437 s
Pillar example:
acl:
- my-filter:
options:
- inet6
terms:
- my-term:
source_port: [1234, 1235]
protocol:
- tcp
- udp
source_address: 1.2.3.4
action: reject
- my-other-term:
source_port:
- [5678, 5680]
protocol: tcp
action: accept
State SLS Example:
{%- set filter_name = 'my-filter' -%}
{%- set my_filter_cfg = salt.netacl.get_filter_pillar(filter_name, pillar_key='firewall') -%}
my_first_filter_state:
netacl.filter:
- filter_name: {{ filter_name }}
- options: {{ my_filter_cfg['options'] | json }}
- terms: {{ my_filter_cfg['terms'] | json }}
- revision_date: false
- revision_no: 5
- debug: true
Or:
my_first_filter_state:
netacl.filter:
- filter_name: my-filter
- merge_pillar: true
- pillar_key: firewall
- revision_date: false
- revision_no: 5
- debug: true
In the example above, as inet6
has been specified in the filter_options
,
the configuration chunk referring to my-term
has been ignored as it referred to
IPv4 only (from source_address
field).
Note
The first method allows the user to eventually apply complex manipulation and / or retrieve the data from external services before passing the data to the state. The second one is more straightforward, for less complex cases when loading the data directly from the pillar is sufficient.
Note
When passing retrieved pillar data into the state file, it is strongly recommended to use the json serializer explicitly (`` | json``), instead of relying on the default Python serializer.
salt.states.netacl.
managed
(name, filters=None, prepend=True, pillar_key='acl', pillarenv=None, saltenv=None, merge_pillar=False, only_lower_merge=False, revision_id=None, revision_no=None, revision_date=True, revision_date_format='%Y/%m/%d', test=False, commit=True, debug=False)¶Manage the whole firewall configuration.
Dictionary of filters for this policy.
If not specified or empty, will try to load the configuration from the pillar,
unless merge_pillar
is set as False
.
True
When merge_pillar
is set as True
, the final list of filters generated by merging
the filters from filters
with those defined in the pillar (if any): new filters are prepended
at the beginning, while existing ones will preserve the position. To add the new filters
at the end of the list, set this argument to False
.
acl
The key in the pillar containing the default attributes values. Default: acl
.
Query the master to generate fresh pillar data on the fly, specifically from the requested pillar environment.
Included only for compatibility with
pillarenv_from_saltenv
, and is otherwise ignored.
False
Merge the filters
will the corresponding values from the pillar. Default: False
.
Note
By default this state does not merge, to avoid any unexpected behaviours.
The merge logic depends on the prepend
argument.
The filters specified through the filters
argument have higher priority
than the pillar.
False
Specify if it should merge only the filters and terms fields. Otherwise it will try
to merge everything at the policy level. Default: False
.
This option requires merge_pillar
, otherwise it is ignored.
False
Dry run? If set as True
, will apply the config, discard and return the changes.
Default: False
and will commit the changes on the device.
Add a comment in the policy config having the description for the changes applied.
The revision count.
True
Boolean flag: display the date when the policy configuration was generated. Default: True
.
%Y/%m/%d
The date format to be used when generating the perforce data. Default: %Y/%m/%d
(<year>/<month>/<day>).
True
Commit? Default: True
.
False
Debug mode. Will insert a new key under the output dictionary,
as loaded_config
containing the raw configuration loaded on the device.
CLI Example:
salt 'edge01.bjm01' state.sls router.acl test=True
Output Example:
edge01.bjm01:
-------------
ID: netacl_example
Function: netacl.managed
Result: None
Comment: Testing mode: Configuration discarded.
Started: 12:03:24.807023
Duration: 5569.453 ms
Changes:
----------
diff:
[edit firewall]
+ family inet {
+ /*
+ ** $Id: netacl_example $
+ ** $Date: 2017/07/03 $
+ ** $Revision: 2 $
+ **
+ */
+ filter my-filter {
+ interface-specific;
+ term my-term {
+ from {
+ source-address {
+ 1.2.3.4/32;
+ }
+ protocol [ tcp udp ];
+ source-port [ 1234 1235 ];
+ }
+ then {
+ reject;
+ }
+ }
+ term my-other-term {
+ from {
+ protocol tcp;
+ source-port 5678-5680;
+ }
+ then accept;
+ }
+ }
+ /*
+ ** $Id: netacl_example $
+ ** $Date: 2017/07/03 $
+ ** $Revision: 2 $
+ **
+ */
+ filter block-icmp {
+ interface-specific;
+ term first-term {
+ from {
+ protocol icmp;
+ }
+ then {
+ reject;
+ }
+ }
+ }
+ }
loaded:
firewall {
family inet {
replace:
/*
** $Id: netacl_example $
** $Date: 2017/07/03 $
** $Revision: 2 $
**
*/
filter my-filter {
interface-specific;
term my-term {
from {
source-address {
1.2.3.4/32;
}
protocol [ tcp udp ];
source-port [ 1234 1235 ];
}
then {
reject;
}
}
term my-other-term {
from {
protocol tcp;
source-port 5678-5680;
}
then accept;
}
}
}
}
firewall {
family inet {
replace:
/*
** $Id: netacl_example $
** $Date: 2017/07/03 $
** $Revision: 2 $
**
*/
filter block-icmp {
interface-specific;
term first-term {
from {
protocol icmp;
}
then {
reject;
}
}
}
}
}
Summary for edge01.bjm01
------------
Succeeded: 1 (unchanged=1, changed=1)
Failed: 0
------------
Total states run: 1
Total run time: 5.569 s
The policy configuration has been loaded from the pillar, having the following structure:
firewall:
- my-filter:
terms:
- my-term:
source_port: [1234, 1235]
protocol:
- tcp
- udp
source_address: 1.2.3.4
action: reject
- my-other-term:
source_port:
- [5678, 5680]
protocol: tcp
action: accept
- block-icmp:
terms:
- first-term:
protocol:
- icmp
action: reject
Example SLS file:
{%- set fw_filters = pillar.get('firewall', {}) -%}
netacl_example:
netacl.managed:
- filters: {{ fw_filters | json }}
- revision_no: 2
- debug: true
Or:
netacl_example:
netacl.managed:
- pillar_key: firewall
- merge_pillar: true
- revision_no: 2
- debug: true
Note
The first method allows the user to eventually apply complex manipulation and / or retrieve the data from external services before passing the data to the state. The second one is more straightforward, for less complex cases when loading the data directly from the pillar is sufficient.
Note
When passing retrieved pillar data into the state file, it is strongly recommended to use the json serializer explicitly (`` | json``), instead of relying on the default Python serializer.
salt.states.netacl.
term
(name, filter_name, term_name, filter_options=None, pillar_key='acl', pillarenv=None, saltenv=None, merge_pillar=False, revision_id=None, revision_no=None, revision_date=True, revision_date_format='%Y/%m/%d', test=False, commit=True, debug=False, source_service=None, destination_service=None, **term_fields)¶Manage the configuration of a specific policy term.
The name of the policy filter.
The name of the term.
Additional filter options. These options are platform-specific. See the complete list of options.
acl
The key in the pillar containing the default attributes values. Default: acl
.
Query the master to generate fresh pillar data on the fly, specifically from the requested pillar environment.
Included only for compatibility with
pillarenv_from_saltenv
, and is otherwise ignored.
False
Merge the CLI variables with the pillar. Default: False
.
The properties specified through the state arguments have higher priority than the pillar.
Add a comment in the term config having the description for the changes applied.
The revision count.
True
Boolean flag: display the date when the term configuration was generated. Default: True
.
%Y/%m/%d
The date format to be used when generating the perforce data. Default: %Y/%m/%d
(<year>/<month>/<day>).
False
Dry run? If set as True
, will apply the config, discard and return the changes.
Default: False
and will commit the changes on the device.
True
Commit? Default: True
.
False
Debug mode. Will insert a new key under the output dictionary,
as loaded_config
containing the raw configuration loaded on the device.
A special service to choose from. This is a helper so the user is able to select a source just using the name, instead of specifying a source_port and protocol.
As this module is available on Unix platforms only, it reads the IANA port assignment from /etc/services.
If the user requires additional shortcuts to be referenced, they can add entries under /etc/services,
which can be managed using the file state
.
A special service to choose from. This is a helper so the user is able to
select a source just using the name, instead of specifying a destination_port and protocol.
Allows the same options as source_service
.
Term attributes. To see what fields are supported, please consult the list of supported keywords. Some platforms have few other optional keywords.
Note
The following fields are accepted:
action
address
address_exclude
comment
counter
expiration
destination_address
destination_address_exclude
destination_port
destination_prefix
forwarding_class
forwarding_class_except
logging
log_name
loss_priority
option
policer
port
precedence
principals
protocol
protocol_except
qos
pan_application
routing_instance
source_address
source_address_exclude
source_port
source_prefix
verbatim
packet_length
fragment_offset
hop_limit
icmp_type
ether_type
traffic_class_count
traffic_type
translated
dscp_set
dscp_match
dscp_except
next_ip
flexible_match_range
source_prefix_except
destination_prefix_except
vpn
source_tag
destination_tag
source_interface
destination_interface
flattened
flattened_addr
flattened_saddr
flattened_daddr
priority
Note
The following fields can be also a single value and a list of values:
action
address
address_exclude
comment
destination_address
destination_address_exclude
destination_port
destination_prefix
forwarding_class
forwarding_class_except
logging
option
port
precedence
principals
protocol
protocol_except
pan_application
source_address
source_address_exclude
source_port
source_prefix
verbatim
icmp_type
ether_type
traffic_type
dscp_match
dscp_except
flexible_match_range
source_prefix_except
destination_prefix_except
source_tag
destination_tag
source_service
destination_service
Example: destination_address
can be either defined as:
destination_address: 172.17.17.1/24
or as a list of destination IP addresses:
destination_address:
- 172.17.17.1/24
- 172.17.19.1/24
or a list of services to be matched:
source_service:
- ntp
- snmp
- ldap
- bgpd
Note
The port fields source_port
and destination_port
can be used as
above to select either a single value, either a list of values, but
also they can select port ranges. Example:
source_port:
- [1000, 2000]
- [3000, 4000]
With the configuration above, the user is able to select the 1000-2000 and 3000-4000 source port ranges.
CLI Example:
salt 'edge01.bjm01' state.sls router.acl
Output Example:
edge01.bjm01:
----------
ID: update_icmp_first_term
Function: netacl.term
Result: None
Comment: Testing mode: Configuration discarded.
Started: 12:49:09.174179
Duration: 5751.882 ms
Changes:
----------
diff:
[edit firewall]
+ family inet {
+ /*
+ ** $Id: update_icmp_first_term $
+ ** $Date: 2017/02/30 $
+ **
+ */
+ filter block-icmp {
+ term first-term {
+ from {
+ protocol icmp;
+ }
+ then {
+ reject;
+ }
+ }
+ }
+ }
Summary for edge01.bjm01
------------
Succeeded: 1 (unchanged=1, changed=1)
Failed: 0
------------
Total states run: 1
Total run time: 5.752 s
Pillar example:
firewall:
- block-icmp:
terms:
- first-term:
protocol:
- icmp
action: reject
State SLS example:
{%- set filter_name = 'block-icmp' -%}
{%- set term_name = 'first-term' -%}
{%- set my_term_cfg = salt.netacl.get_term_pillar(filter_name, term_name) -%}
update_icmp_first_term:
netacl.term:
- filter_name: {{ filter_name }}
- filter_options:
- not-interface-specific
- term_name: {{ term_name }}
- {{ my_term_cfg | json }}
Or directly referencing the pillar keys:
update_icmp_first_term:
netacl.term:
- filter_name: block-icmp
- filter_options:
- not-interface-specific
- term_name: first-term
- merge_pillar: true
Note
The first method allows the user to eventually apply complex manipulation and / or retrieve the data from external services before passing the data to the state. The second one is more straightforward, for less complex cases when loading the data directly from the pillar is sufficient.
Note
When passing retrieved pillar data into the state file, it is strongly recommended to use the json serializer explicitly (`` | json``), instead of relying on the default Python serializer.