| Age | Commit message (Collapse) | Author |
|
T7732: VPP add modprobe vfio modules
|
|
VPP can use 2M hugepages and 1G hugepages at the same time
|
|
Add vfio modules before VPP starting and remove modules after
the VPP configuration deleted
modules:
- vfio
- vfio_iommu_type1
- vfio_pci
|
|
kernel-updates
|
|
- Moved is_interface_has_mac outside of _get_summary_data
- Updated MAC definition in _format_kernel_data to use is_interface_has_mac
|
|
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
|
|
Fixes for the following:
- Tunnel interfaces report an IP address in the field where a MAC is expected
- Podman created veth interfaces list it's parent network as a VRF instead of the actual VRF.
- Long descriptions should wrap
- Non useful interfaces shouldn't be shown.
|
|
|
|
T7678: Move "vpp settings host-resources" to "system option resource-limits"
|
|
T7687: Add op-mode for VLAN-to-VNI mapping
|
|
|
|
kea: T6211: add VRF support for KEA dhcp server
|
|
T7718: expose validate_tree* methods and add validate-config script
|
|
|
|
|
|
|
|
|
|
http-api: T7711: config sync method should be async
|
|
T7716: VPP: Change defaults and add a warning about maximum routes count for current configuration
|
|
current configuration
|
|
This is needed due to the recent change of the underlying method
_configure_op to async.
|
|
It will be configured in section "system option host-resources"
|
|
|
|
|
|
T7697: Merge vyos-vpp repo into vyos-1x
|
|
op-mode: T7583: add a new operational command runner
|
|
vxlan: T7468: add VLAN-to-VNI mapping description
|
|
Ignore import vpp_api to skip pylint check.
The `python3-vpp-api` depends on `vpp` and we should not use VPP
in the container.
Use this workaround to skip pylint check.
|
|
The pylint does not like the IProute import (though it works fine)
************* Module vpp
src/conf_mode/vpp.py:21:0: E0611: No name 'IPRoute' in module 'pyroute2' (no-name-in-module)
Do this change only for pylint
|
|
- Added op-mode commands for VLAN-to-VNI mappings
- Commands added:
- show interfaces vxlan <interface> vlan-to-vni
- show interfaces vxlan <interface> vlan-to-vni detail
- show interfaces vxlan <interface> vlan-to-vni vlan <vlan-id>
- show interfaces vxlan <interface> vlan-to-vni vlan <vlan-id> detail
|
|
|
|
|
|
|
|
|
|
Added check for None value from leaf_node_changed when changing VLAN-to-VNI description.
|
|
T7668: bail out early in get_cli_kernel_options() if nothing is to do
|
|
Commit a99ca6d11 ("op-mode: T7403: add option for forcefully removing a
container image") did not account for the case where a single image ID is
passed as a string. This led to incorrect behavior during iteration.
Without converting the string to a list, the code iterates over individual
characters of the image ID rather than treating it as a single item. As a
result, the system attempts to delete non-existent container images based
on these characters.
|
|
serial: T7484: treat unavailable serial console devices as non-fatal
|
|
op-mode: T7403: add an option to forcefully remove a container image
|
|
T7651: VPP use pid not ppid as default in vpp-failure-handler.service script
|
|
If there is no Kernel option to pass, there is no need to run the code to
completion. Bail out early.
|
|
* VPP: T7175
Added conf mode CLI for VPP sflow plugin and updated VPP template to include plugin.
* VPP: T7175: Conf mode CLI and startup template.
* T7175: VPP fix sFlow verify use vpp enstead of enable_vpp
* T7175: VPP add sFlow smoketest
* T7175: VPP remove unused config_changed variable
---------
Co-authored-by: Viacheslav <v.gletenko@vyos.io>
|
|
Instead of failing or starting agetty on an invalid device, we now check
whether the specified serial console device is a valid TTY. If it's not, we
display a warning to the user and skip starting the agetty process.
|
|
A subsequent call to restart_login_consoles() will always call "systemctl
daemon-reload" and thus we can spare one call.
|
|
If we do not have the system options kernel the update fails
Change order for variables `k_cpu_opts` and `k_memory_opts`
If we do not have kernel_options set empty dictionary as
NoneType object has no attribue `get`
```
>>> kernel_options=None
>>> kernel_options.get('cpu', {})
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: 'NoneType' object has no attribute 'get'
>>>
>>>
>>> kernel_options={}
>>> kernel_options.get('cpu', {})
{}
>>>
```
|
|
If you have multiple images (tags) loaded for a container which point to the
exact same hash - you can only forcefully delete them.
vyos@vyos:~$ delete container image 70dc5806
Error: unable to delete image "70dc5806" by ID with more than one tag
([registry.io/foo/bar:v1.0 registry.io/foo/bar:v1.0.1]):
please force removal
In additon an image that is still beeing used can not be removed from the
system as this would cause infinite container restarts.
vyos@vyos:~$ delete container image all force
Cannot delete image "d205499ae8bb" because it is currently being used by
container "3b290f90e83d"!
|
|
|
|
The reset_section.py script is used in a limited case for both (1)
recovery from an apply error in a config mode script and (2) recovery by
a service crash in a failure handler service. In the latter case, the
default pid must be os.getpid, not os.getppid, so as not to be called by
pid 1.
|
|
T7628: Fix non-TPM backed config encryption
|
|
qat: T7662: add PCI ID range for Intel C62x virtual function devices
|