| Age | Commit message (Collapse) | Author |
|---|
|
After e7bb65894 ("vrf: T6189: render FRR L3VNI configuration when creating VRF
instance") we need to ensure that the VRF L3VNI configuration is removed in FRR
prior to removing the BGP VRF instance.
The reason is [1] where FRR only allows VRF BGP instance to be removed when
there is NO VNI configured anymore.
1: https://github.com/FRRouting/frr/blob/064c3494527b9e84260410006768ed38e57e1de7/bgpd/bgp_vty.c#L1646-L1650
|
|
When adding and removing VRF instances on the fly it was noticed that the vni
statement under the VRF instance in FRR vanishes. This was caused by a race
condition which was previously designed to fix another bug.
The wierd design of a Python helper below the VRF tree to only generate the
VNI configuration nodes is now gone and all is rendered in the proper place.
|
|
|
|
T6221: Return default ip rule values after deleting VRF
|
|
|
|
Fix for restoring default ip rule values after deleting VRF
Defult values:
```
$ ip rule
0: from all lookup local
32766: from all lookup main
32767: from all lookup default
```
After adding and deleting a VRF we get unexpected values:
```
$ ip rule
1000: from all lookup [l3mdev-table]
2000: from all lookup [l3mdev-table] unreachable
32765: from all lookup local
32766: from all lookup main
32767: from all lookup default
```
|
|
T6242: load-balancing reverse-proxy: Ability for ssl backends to not verify server certificates
|
|
server certificates
|
|
pki: T6241: Fix dependency updates on PKI changes
|
|
T5535: firewall: migrate command <set system ip disable-directed-broadcast> to firewall global-optinos
|
|
|
|
to firewall global-optinos
|
|
Check if DH is configured for OpenVPN but does not exist in the
PKI section
```
set pki dh dh-correct parameters 'xxxx'
set interfaces openvpn vtun10 tls dh-params 'dh-fake'
File "/usr/libexec/vyos/conf_mode/interfaces_openvpn.py", line 208, in verify_pki
pki_dh = pki['dh'][tls['dh_params']]
~~~~~~~~~^^^^^^^^^^^^^^^^^^
KeyError: 'dh-fake'
```
|
|
fails (#3296)
|
|
T5169: Add PoC for generating CGNAT rules rfc6888
|
|
|
|
Add PoC for generating CGNAT rules
https://datatracker.ietf.org/doc/html/rfc6888
Not all requirements are implemented, but some of them.
Implemented:
REQ-2
```
A CGN MUST have a default "IP address pooling" behavior of "Paired"
CGN must use the same external IP
address mapping for all sessions associated with the same internal
IP address, be they TCP, UDP, ICMP, something else, or a mix of
different protocols.
```
REQ-3
```
The CGN function SHOULD NOT have any limitations on the size
or the contiguity of the external address pool
```
REQ-4
```
A CGN MUST support limiting the number of external ports (or,
equivalently, "identifiers" for ICMP) that are assigned per
subscriber
```
CLI:
```
set nat cgnat pool external ext1 external-port-range '1024-65535'
set nat cgnat pool external ext1 per-user-limit port '1000'
set nat cgnat pool external ext1 range 192.0.2.222/32
set nat cgnat pool internal int1 range '100.64.0.0/28'
set nat cgnat rule 10 source pool 'int1'
set nat cgnat rule 10 translation pool 'ext1'
```
|
|
|
|
ethernet: T5862: default MTU is not acceptable in some environments
|
|
login: T5875: fix corner case for KeyError: 'getpwuid(): uid not found: XXXX'
|
|
|
|
'upper'
Commit b30faa43c (container: T6208: rename "cap-add" CLI node to "capability")
added an AttributeError referencing an out of scope variable.
This has been fixed.
|
|
Commit 1b364428f ("login: T5875: restore home directory permissions only when
needed") added logic to chown the users home directory if it's UID changes.
This might happen when a user account is deleted and re-added to the system.
Under rar e circumstances it was possible that the implementation triggered
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
KeyError: 'getpwuid(): uid not found: XXXX'
This has been fixed by re-arranging the code path with an additional try/except
if the PW database information could not be retrieved leading to an implicit
chown() of the home directory to the user beeing added.
|
|
Containers have the ability to add Linux system capabilities to them, this is
done using the "set container name <name> cap-add" command.
The CLI node sounds off and rather should be "set container name <name>
capability" instead as we use and pass a capability to a container and not
add/invent new ones.
|
|
mac node
The problem was introduced in [1] but the config migrator part unfortunately
was added to the wrong version [2]. As IPoE config version 0 was only active
during the 1.3 development cycle and VyOS 1.3.0 was already released with config
version 1 we can safely drop the migrator 0-to-1 and move the code to 1-to-2 to
properly support upgrades from VyOS 1.3 -> 1.4 or newer.
1: https://github.com/vyos/vyos-1x/commit/05df2a5f021f0c7aab7c06db645d210858b6e98d#diff-08291bf77870abe3af8bbe3e8ce4bbf344fd0498b2c5c75a75aa7235d381c88eL168
2: https://github.com/vyos/vyos-1x/commit/05df2a5f021f0c7aab7c06db645d210858b6e98d#diff-b8bb58b75607d3653e74d82eff02442f9f3ab82698f160ba37858f7cdf6c79ccR44-R46
|
|
There are cloud environments available where the maximum supported ethernet
MTU is e.g. 1450 bytes, thus we clamp this to the adapters maximum MTU value
or 1500 bytes - whatever is lower.
|
|
bgp: T5943: BGP Peer-group members must be all internal or all external
|
|
T6068: dhcp-server: add command <set service dhcp-server high-availability mode>
|
|
|
|
mode> so user can define what type of ha use: active-active or active-passive
|
|
found using "git ls-files *.py | xargs pylint | grep W0611"
|
|
* Use interface_exists() outside of verify()
* Use verify_interface_exists() in verify() to drop common error message
|
|
configverify: T6198: add common helper for PKI certificate validation
|
|
The next evolutional step after adding get_config_dict(..., with_pki=True) is
to add a common verification function for the recurring task of validating SSL
certificate existance in e.g. EAPoL, OpenConnect, SSTP or HTTPS.
|
|
T6192: allow binding SSH to multiple VRF instances
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Currently VyOS only supports binding a service to one individual VRF. It might
become handy to have the services (initially it will be VRF, NTP and SNMP) be
bound to multiple VRFs.
Changed VRF from leafNode to multi leafNode with defaultValue: default - which
is the name of the default VRF.
|
|
|
|
system: T6193: invalid warning "is not a DHCP interface but uses DHCP name-server option"
|
|
T6178: Check that certificate exists during reverse-proxy commit
|
|
name-server option"
This fixes an invalid warning when using a DHCP VLAN interface to retrieve the
system nameserver to be used. VLAN CLI config is not properly expanded
leading to a false warning:
[ system name-server eth1.10 ]
WARNING: "eth1.10" is not a DHCP interface but uses DHCP name-server option!
|
|
|
|
dhcp-server: T4718: Listen-address is not commit if the ip address is on the interface with vrf
|
|
peer-group
changed exception condition
Improved route_reflector_client test
|
|
ipsec: T5606: T5871: Use multi node for CA certificates
|
