You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: document/docs/troubleshooting/troubleshooting.md
+46-12Lines changed: 46 additions & 12 deletions
Original file line number
Diff line number
Diff line change
@@ -89,7 +89,25 @@ There are two common root causes.
89
89
2. Network interfaces are not configured as expected.
90
90
* Check [multi-nicd log](#get-multi-nicd-log).
91
91
92
-
- If getting `cannot list address on <SECONDARY INTERFACE>`, please confirm whether IPv4 address on the host.
92
+
- If getting `cannot list address on <SECONDARY INTERFACE>`, please confirm whether IPv4 address on the host.
93
+
- If getting `cannot get PCI info: Get "https://pci-ids.ucw.cz/v2.2/pci.ids.gz": net/http: TLS handshake timeout`, some environment variables need to be set in the config for the multi-nicd container to reach the above address via proxy settings.
94
+
95
+
apiVersion: multinic.fms.io/v1
96
+
kind: Config
97
+
metadata:
98
+
name: multi-nicd
99
+
...
100
+
spec:
101
+
...
102
+
daemon:
103
+
env:
104
+
- name: HTTP_PROXY
105
+
value: <REPLACE WITH YOUR HTTPS_PROXY>
106
+
- name: HTTPS_PROXY
107
+
value: <REPLACE WITH YOUR HTTPS_PROXY>
108
+
- name: NO_PROXY
109
+
value: <REPLACE WITH YOUR NO_PROXY>
110
+
93
111
- Otherwise, please refer to [check interfaces at node's host network](#check-host-secondary-interfaces).
94
112
95
113
### Pod failed to start
@@ -222,20 +240,36 @@ Zero config occurs when CNI cannot generate configurations from the network-atta
222
240
### Ping failed
223
241
**Issue:** Pods cannot ping each other.
224
242
225
-
Check route status in multinicnetworks.multinic.fms.io.
226
-
```bash
227
-
kubectl get multinicnetwork.multinic.fms.io multinic-ipvlanl3 -o json\
228
-
| jq -r .status.routeStatus
229
-
```
243
+
* If the CNI operates at Layer 2 (such as MACVLAN or IPVLAN with L2), please confirm whether the defined Pod CIDR is routable within your cluster.
244
+
245
+
For bare metal cluster which has only a certain VLAN range opened on the switch,
246
+
please define a VLAN interface instead of the physical NIC on the node.
247
+
Usually for a bare metal node with a secondary interface,
248
+
the two ports of NIC2 will be defined as tenant-bond for redundancy,
249
+
the VLAN interface should be defined following the naming vlanXXX@tenant-bond,
250
+
where XXX represents a valid open VLAN ID.
251
+
252
+
Please see the following example:
253
+
254
+
13769: tenant-bond: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP group
* If the CNI operates at Layer 3, check route status in `multinicnetworks.multinic.fms.io`.
260
+
261
+
kubectl get multinicnetwork.multinic.fms.io multinic-ipvlanl3 -o json\
262
+
| jq -r .status.routeStatus
263
+
264
+
-*WaitForRoutes*: the new cidr is just recomputed and waiting for route update.
265
+
-*Failed*: some route cannot be applied, need attention. Check [multi-nicd log](#get-multi-nicd-log)
266
+
-*Unknown*: some daemon cannot be connected.
267
+
-*N/A*: there is no L3 configuration applied. Check whether multinicnetwork.multinic.fms.io is defined with L3 mode and cidrs.multinic.fms.io is created.
230
268
231
-
-*WaitForRoutes*: the new cidr is just recomputed and waiting for route update.
232
-
-*Failed*: some route cannot be applied, need attention. Check [multi-nicd log](#get-multi-nicd-log)
233
-
-*Unknown*: some daemon cannot be connected.
234
-
-*N/A*: there is no L3 configuration applied. Check whether multinicnetwork.multinic.fms.io is defined with L3 mode and cidrs.multinic.fms.io is created.
269
+
kubectl get cidrs.multinic.fms.io
235
270
236
-
kubectl get cidrs.multinic.fms.io
271
+
-*Success*: check [set required security group rules](#set-security-groups)
237
272
238
-
-*Success*: check [set required security group rules](#set-security-groups)
239
273
### TCP/UDP communication failed.
240
274
**Issue:** Pods can ping each other but do not get response from TCP/UDP communication such as iPerf.
- [SR-IoV with Multi-NIC IPAM](https://github.com/foundation-model-stack/multi-nic-cni/tree/main/config/samples/multinicnetwork/sriov.yaml) ( from v1.2.0 )
168
+
- [SR-IoV with Multi-NIC IPAM](https://github.com/foundation-model-stack/multi-nic-cni/tree/main/config/samples/multinicnetwork/sriov.yaml)
- [Mellanox Host Device without IPAM](https://github.com/foundation-model-stack/multi-nic-cni/tree/main/config/samples/multinicnetwork/mellanox_hostdevice.yaml) ( from v1.2.0 )
172
+
- [Mellanox Host Device with Host Device IPAM](https://github.com/foundation-model-stack/multi-nic-cni/tree/main/config/samples/multinicnetwork/mellanox_hostdevice.yaml)
- [IPVLAN L2 with unmanaged HostInterface](https://github.com/foundation-model-stack/multi-nic-cni/tree/main/config/samples/multinicnetwork/ipvlanl2_unmanaged.yaml) ( from v1.2.1 )
- [IPVLAN L2 with AWS-VPC-connecting IPAM](https://github.com/foundation-model-stack/multi-nic-cni/tree/main/config/samples/multinicnetwork/awsipvlan.yaml) ( from v1.1.0 )
177
+
- [IPVLAN L2 with AWS-VPC-connecting IPAM](https://github.com/foundation-model-stack/multi-nic-cni/tree/main/config/samples/multinicnetwork/awsipvlan.yaml)
0 commit comments