Skip to content

Commit d0068f0

Browse files
ehsjoarJonas Arndtranjinimn
authored
Adding Bond CNI Documentation Edge 1378 (suse-edge#747)
* First take on adding bond cni information * Minor updates * Minor updates and removed a pdf file that was accidentally added * Minor updates from Alberto's review * Added CNI info for Calico too * Changes to address review comments * Fixed URL reference * Update asciidoc/product/atip-features.adoc Co-authored-by: ranjinimn <[email protected]> --------- Co-authored-by: Jonas Arndt <[email protected]> Co-authored-by: ranjinimn <[email protected]>
1 parent d3cf36d commit d0068f0

File tree

1 file changed

+302
-2
lines changed

1 file changed

+302
-2
lines changed

‎asciidoc/product/atip-features.adoc‎

Lines changed: 302 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -234,7 +234,8 @@ The script is available at https://raw.githubusercontent.com/suse-edge/atip/refs
234234

235235
==== Cilium
236236
`Cilium` is the default CNI plug-in for SUSE Edge for Telco.
237-
To enable Cilium on RKE2 cluster as the default plug-in, the following configurations are required in the `/etc/rancher/rke2/config.yaml` file:
237+
To enable Cilium on RKE2 cluster as the default plug-in, the following
238+
configuration is required in the `/etc/rancher/rke2/config.yaml` file:
238239

239240
[,yaml]
240241
----
@@ -253,11 +254,310 @@ cni:
253254
- cilium
254255
----
255256

257+
==== Calico
258+
`Calico` is another CNI plug-in for SUSE Edge for Telco.
259+
To enable Calico on RKE2 cluster as the default plug-in, the following
260+
configuration is required in the `/etc/rancher/rke2/config.yaml` file:
261+
262+
[,yaml]
263+
----
264+
cni:
265+
- calico
266+
----
267+
268+
This can also be specified with command-line arguments, that is, `--cni=calico` into the server line in `/etc/systemd/system/rke2-server` file.
269+
270+
To use the `SR-IOV` network operator described in the xref:option2-sriov-helm[next section], use `Multus` with another CNI plug-in, like `Cilium` or `Calico`, as a secondary plug-in.
271+
272+
[,yaml]
273+
----
274+
cni:
275+
- multus
276+
- calico
277+
----
278+
256279
[NOTE]
257280
====
258-
For more information about CNI plug-ins, visit https://docs.rke2.io/install/network_options[Network Options].
281+
For more information about CNI plug-ins, see https://docs.rke2.io/install/network_options[Network Options].
259282
====
260283

284+
==== Bond CNI
285+
In general terms, bonding provides a method for aggregating multiple network
286+
interfaces into a single logical "bonded" interface. This is typically used to
287+
increase service availability by introducing redundant networking paths, but can
288+
also be used to increase bandwidth with certain bond modes.
289+
The following CNI plug-ins are supported for the Bond CNI plugin in combination with multus:
290+
291+
* MACVLAN
292+
* Host Device
293+
* SR-IOV
294+
295+
===== Bond CNI with MACVLAN
296+
To use the Bond CNI plugin with MACVLAN two free interfaces are needed in the
297+
container. The following example uses 'enp8s0' and 'enp9s0'. Start by creating network
298+
attachment definitions for them:
299+
300+
*NetworkAttachmentDefinition enp8s0*
301+
302+
[,shell,subs="attributes"]
303+
----
304+
apiVersion: "k8s.cni.cncf.io/v1"
305+
kind: NetworkAttachmentDefinition
306+
metadata:
307+
name: enp8s0-conf
308+
spec:
309+
config: '{
310+
"cniVersion": "0.3.1",
311+
"plugins": [
312+
{
313+
"type": "macvlan",
314+
"capabilities": { "ips": true },
315+
"master": "enp8s0",
316+
"mode": "bridge",
317+
"ipam": {}
318+
}, {
319+
"capabilities": { "mac": true },
320+
"type": "tuning"
321+
}
322+
]
323+
}'
324+
----
325+
326+
*NetworkAttachmentDefinition enp9s0*
327+
328+
[,shell,subs="attributes"]
329+
----
330+
apiVersion: "k8s.cni.cncf.io/v1"
331+
kind: NetworkAttachmentDefinition
332+
metadata:
333+
name: enp9s0-conf
334+
spec:
335+
config: '{
336+
"cniVersion": "0.3.1",
337+
"plugins": [
338+
{
339+
"type": "macvlan",
340+
"capabilities": { "ips": true },
341+
"master": "enp9s0",
342+
"mode": "bridge",
343+
"ipam": {}
344+
}, {
345+
"capabilities": { "mac": true },
346+
"type": "tuning"
347+
}
348+
]
349+
}'
350+
----
351+
352+
After this, add a network attachment definition for the bond itself.
353+
354+
*NetworkAttachmentDefinition bond*
355+
356+
[,shell,subs="attributes"]
357+
----
358+
apiVersion: "k8s.cni.cncf.io/v1"
359+
kind: NetworkAttachmentDefinition
360+
metadata:
361+
name: bond-net1
362+
spec:
363+
config: '{
364+
"type": "bond",
365+
"cniVersion": "0.3.1",
366+
"name": "bond-net1",
367+
"mode": "active-backup",
368+
"failOverMac": 1,
369+
"linksInContainer": true,
370+
"miimon": "100",
371+
"mtu": 1500,
372+
"links": [
373+
{"name": "net1"},
374+
{"name": "net2"}
375+
],
376+
"ipam": {
377+
"type": "static",
378+
"addresses": [
379+
{
380+
"address": "192.168.200.100/24",
381+
"gateway": "192.168.200.1"
382+
}
383+
],
384+
"subnet": "192.168.200.0/24",
385+
"routes": [{
386+
"dst": "0.0.0.0/0"
387+
}]
388+
}
389+
}'
390+
----
391+
392+
The IP address assignment here is static and defines the address of the bond as
393+
'192.168.200.100' on a /24 network, with a gateway residing on the network's first
394+
available address. In the bond's network attachment we also define the type of
395+
bond we want. In this case it is active-backup.
396+
397+
To use this bond, the pod needs to know about all interfaces. An example pod
398+
definition might look like this:
399+
400+
[,shell,subs="attributes"]
401+
----
402+
apiVersion: v1
403+
kind: Pod
404+
metadata:
405+
name: test-pod
406+
annotations:
407+
k8s.v1.cni.cncf.io/networks: '[
408+
{"name": "enp8s0-conf",
409+
"interface": "net1"
410+
},
411+
{"name": "enp9s0-conf",
412+
"interface": "net2"
413+
},
414+
{"name": "bond-net1",
415+
"interface": "bond0"
416+
}]'
417+
spec:
418+
restartPolicy: Never
419+
containers:
420+
- name: bond-test
421+
image: alpine:latest
422+
command:
423+
- /bin/sh
424+
- "-c"
425+
- "sleep 60m"
426+
imagePullPolicy: IfNotPresent
427+
----
428+
429+
Note how the annotation refers to all networks and how it defines the mapping
430+
between the interfaces 'enp8s0 -> net1', and 'enp9s0->net2'.
431+
432+
===== Bond CNI with Host Device
433+
To use the Bond CNI plugin with host device, two free interfaces are needed on
434+
the host. These interfaces are then mapped through to the container. The
435+
following example uses 'enp8s0' and 'enp9s0'. Start by creating network
436+
attachment definitions for them:
437+
438+
*NetworkAttachmentDefinition enp8s0*
439+
440+
[,shell,subs="attributes"]
441+
----
442+
apiVersion: "k8s.cni.cncf.io/v1"
443+
kind: NetworkAttachmentDefinition
444+
metadata:
445+
name: enp8s0-hostdev
446+
spec:
447+
config: '{
448+
"cniVersion": "0.3.1",
449+
"plugins": [
450+
{
451+
"type": "host-device",
452+
"name": "host0",
453+
"device": "enp8s0",
454+
"ipam": {}
455+
}]
456+
}'
457+
----
458+
459+
*NetworkAttachmentDefinition enp9s0*
460+
461+
[,shell,subs="attributes"]
462+
----
463+
apiVersion: "k8s.cni.cncf.io/v1"
464+
kind: NetworkAttachmentDefinition
465+
metadata:
466+
name: enp9s0-hostdev
467+
spec:
468+
config: '{
469+
"cniVersion": "0.3.1",
470+
"plugins": [
471+
{
472+
"type": "host-device",
473+
"name": "host0",
474+
"device": "enp9s0",
475+
"ipam": {}
476+
}]
477+
}'
478+
----
479+
480+
After this, add network attachment definition for the bond itself. This is
481+
similar to the MACVLAN use case.
482+
483+
*NetworkAttachmentDefinition bond*
484+
485+
[,shell,subs="attributes"]
486+
----
487+
apiVersion: "k8s.cni.cncf.io/v1"
488+
kind: NetworkAttachmentDefinition
489+
metadata:
490+
name: bond-net1
491+
spec:
492+
config: '{
493+
"type": "bond",
494+
"cniVersion": "0.3.1",
495+
"name": "bond-net1",
496+
"mode": "active-backup",
497+
"failOverMac": 1,
498+
"linksInContainer": true,
499+
"miimon": "100",
500+
"mtu": 1500,
501+
"links": [
502+
{"name": "net1"},
503+
{"name": "net2"}
504+
],
505+
"ipam": {
506+
"type": "static",
507+
"addresses": [
508+
{
509+
"address": "192.168.200.100/24",
510+
"gateway": "192.168.200.1"
511+
}
512+
],
513+
"subnet": "192.168.200.0/24",
514+
"routes": [{
515+
"dst": "0.0.0.0/0"
516+
}]
517+
}
518+
}'
519+
----
520+
521+
The IP address assignment here is static and defines the address of the bond as
522+
'192.168.200.100' on a /24 network, with a gateway residing on the network's
523+
first available address. In the bond's network attachment, define the type of
524+
bond. In this case it is active-backup.
525+
526+
To use this bond, the pod needs to know about all interfaces. An example pod
527+
definition for bond with host devices might look like this:
528+
529+
[,shell,subs="attributes"]
530+
----
531+
apiVersion: v1
532+
kind: Pod
533+
metadata:
534+
name: test-pod
535+
annotations:
536+
k8s.v1.cni.cncf.io/networks: '[
537+
{"name": "enp8s0-hostdev",
538+
"interface": "net1"
539+
},
540+
{"name": "enp9s0-hostdev",
541+
"interface": "net2"
542+
},
543+
{"name": "bond-net1",
544+
"interface": "bond0"
545+
}]'
546+
spec:
547+
restartPolicy: Never
548+
containers:
549+
- name: bond-test
550+
image: alpine:latest
551+
command:
552+
- /bin/sh
553+
- "-c"
554+
- "sleep 60m"
555+
imagePullPolicy: IfNotPresent
556+
----
557+
558+
===== Bond CNI with SR-IOV
559+
Using the Bond CNI with SR-IOV is fairly straight forward. For more details on how to set up SR-IOV, see <<sriov>>. As described there, you have to create `SriovNetworkNodePolicies` that defines `resourceNames`, as well as number of virtual functions and such. The `resourceNames` are being used by the `SriovNetwork` which is used as interfaces in the pod definition. The bond definition is exactly the same as for the MACVLAN and host device cases.
560+
261561
[#sriov]
262562
=== SR-IOV
263563

0 commit comments

Comments
 (0)