Age | Commit message (Collapse) | Author |
|
The current VyOS CLI parser code written in Python contains a ton of duplicates
which I can also hold myself accountable for - or maybe mainly me - depends on
the angle of judge.
While providing a new update() method in vyos.ifconfig.interfaces() this is
extended for pdeudo-ethernet interfaces in the derived class.
|
|
The current VyOS CLI parser code written in Python contains a ton of duplicates
which I can also hold myself accountable for - or maybe mainly me - depends on
the angle of judge.
While providing a new update() method in vyos.ifconfig.interfaces() this is
extended for bridge interfaces in the derived bridge class.
Signed-off-by: Christian Poessinger <christian@poessinger.com>
|
|
The current VyOS CLI parser code written in Python contains a ton of duplicates
which I can also hold myself accountable for - or maybe mainly me - depends on
the angle of judge.
While providing a new update() method in vyos.ifconfig.interfaces() this is
extended for ethernet based interfaces which also supports 802.1q, 802.1ad
VLANs. This commit migrates the existing codebase for an ethernet based
interfaces and implements the missing parts for VLANs. Adding or migrating other
interfaces (e.g. bridge or bond) will become much easier as they must reuse
the entire functionality - we now walk towards a single codepath.
Thanks for all who made this combined effort possible!
Signed-off-by: Christian Poessinger <christian@poessinger.com>
|
|
l2tpv3, wireguard, wirelessmodem, nat all require additional Kernel modules
to be present on the system. Each and every interface implemented their own
way of loading a module - by copying code.
Use a generic function, vyos.util.check_kmod() to load any arbitrary kernel
module passed as string or list.
|
|
Since it's relatively rarely used, Python's startup time should't
be much of a problem.
|
|
|
|
|
|
The only calls to config (return_value, return_values, exists) do not
throw VyOSError; remove unneeded except.
|
|
Fix for https://phabricator.vyos.net/T2725
T2492 / a07e22377ab83104ac925e13d1824f241f0f8d4a
introduced a change which broke the initialization of
the user dict. In case the config contained an user
without an encrypted-password set, the property would
be missing and the commit would crash with
`KeyError: 'password_encrypted'`
|
|
|
|
The calculated size of DHCP server address pools was not corrent. The
fact that both boundaries of address ranges are inclusive wasn't
accounted for, so the calculated size was too small by 1 address
per range.
|
|
nat: T2709: remove 'translation address' mandatory check
|
|
Rules without a translation address are also valid, they'll modify just
the port and leave the address intact.
This also used to be a valid syntax and it caused an error on upgrade.
|
|
Improve readability of the port validation. Also fixes T2708.
|
|
Add additional verify() stage to check that the relaying interface actually
exists on the system.
|
|
Since it is not impossible to change MACVLAN interface mode on the fly, this fix adds a new one trigger to recreate it from scratch if the mode changed in system config
|
|
|
|
If flow records contain entries with skipped details, this does not allow preparing it for the output table. This fix use safe .get() function to return empty values instead crashing.
Also, added several small formatting fixes.
|
|
Please note that either interface or address can be configured.
system {
options {
http-client {
source-interface eth0
source-address 1.1.1.1
}
}
}
|
|
The 3rd party library used for calculating the SNMP hashes in advance only
worked for SHA and nod for MD5 as SHA was hardcoded [1]. The code has been
replaced by a class-less implementation providing only the required
functionality.
[1]: https://github.com/TheMysteriousX/SNMPv3-Hash-Generator/issues/2
|
|
|
|
As of now when adding new credentials for any SNMPv3 user we submit the
credential either plaintext or encrypted. A plaintext credential will be hashed
by SNMPd in the background and then passed back into the CLI so it's not stored
in cleartext. This feels like the wrong way in changing the CLI content with
data produced by a 3rd party daemon which implements the service.
It feels like the tail wiggles the entire dog.
This should be changed in the following way:
- After retrieving the plaintext password from CLI, use Python to hash the key
in advance
- Re-populate the encrypted key into the CLI and drop the plaintext one
- Generate service configuration and continue startup of SNMPd
This also fixes a race condition when SNMPd started up but not properly
provided the hasehd keys in the configuration resulting in a ConfigurationError.
Now as we also support binding SNMPd to a VRF this fixes a deadlock situation
on bootup as we can only bind late to the VRF and require up to 5 restarts of
the service - but the service will never start.
|
|
When migrating the conf from VyOS 1.2 to 1.3 a configuration error could appear
if the user specified "info" as loglevel instead of "INFO". There was no input
validation done in 1.2 but this is now enforced in 1.3.
In VyOS 1.3 loglevel will be always lowercase on the CLI and when migrating the
config this is transformed. Also VyOS 1.2 accpeted any arbitrary loglevel. If
an invalid loglevel is found it will be set to info.
|
|
Linux tries to bind sshd to the VRF but it is yet not ready - for any arbitrary
reason. After restarting SSH to often (rate-limiting) it is blocked by systemd.
Using Restart/RestartSec is not enough - systemd services use start rate
limiting (enabled by default). If service is started more than StartLimitBurst
times in StartLimitIntervalSec seconds is it not permitted to start any more.
Parameters are inherited from DefaultStartLimitIntervalSec (default 10s) and
DefaultStartLimitBurst (default 5).
|
|
|
|
|
|
ids: T2659: Implement fastnetmon support
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ifconfig: T2653: move macsec interface to get_config_dict()
|
|
|
|
|
|
|
|
While moving towards a general interface abstraction based on get_config_dict()
and the use of vyos.ifconfig.Interfaces().update() it also makes sense, to
split out common verification code to a common util file - instead of
duplicating the code, which is infact one of the main forces drivind this
transition.
vyos.configverify will hold common functions called via verify() from our
src/conf_mode scripts so we do not need to copy/paste general verifications
methods.
|
|
By implementation misstake we exited the migrationscript if no VXLAN interface
was found but this was wrong as the same loop is used on pseudo-ethernet
interfaces. The Migrator previously only worked on pseudo-ethernet when
also a VXLAN interface was present. This has been corrected.
|
|
|
|
This changes the dummy interface implementation to make use of get_config_dict()
and also implement a new vyos.ifconfig.Interface().update() function to gather
all the scattered calls to update common interface configuration options.
Derived classes of Interface() should extend update() to their needs for their
special interface type - e.g. bond or bridge.
|
|
|
|
|
|
Jinja template contains some workarounds like {% if port is string %}, this
depends of the resolution of https://phabricator.vyos.net/T2636
|