Skip to content

Commit

Permalink
Merge pull request #1325 from spidernet-io/fix/egres-e2e
Browse files Browse the repository at this point in the history 
Enhance code fix daily e2e failed
  • Loading branch information
@weizhoublue
weizhoublue authored Apr 25, 2024
2 parents 0a99fa4 + f9a36df commit f188111
Show file tree
Hide file tree
Showing 7 changed files with 95 additions and 26 deletions.
2 changes: 1 addition & 1 deletion .github/workflows/call-e2e.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -124,7 +124,7 @@ jobs:
- name: Setup Kind Cluster
uses: nick-invision/retry@v3
with:
timeout_minutes: 10
timeout_minutes: 20
max_attempts: 1
shell: bash
command: |
Expand Down
4 changes: 2 additions & 2 deletions test/doc/reliability.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,8 @@

| Case ID | Title | Priority | Smoke | Status | Other |
|---------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|----------|---------|--------|-------|
| R00001 | Use `kwok` to create 10 `Node`, create `Deployment` with 1000 replicas, create `Policy` and set `PodSelector` to match `Deployment`,<br/>all matched `Pod`'s egress IP in the real node is `eip` | p3 | false | done | |
| R00002 | Use `kwok` to create 10 `Node`, create `Deployment` with 1000 replicas, create `Policy` and set `PodSelector` to match `Deployment`,<br/>After restarting `Deployment` successfully, all matched `Pod`'s egress IP in the real node is `eip` | p3 | false | done | |
| R00001 | Use `kwok` to create 10 `Node`, create `Deployment` with 1000 replicas, create `Policy` and set `PodSelector` to match `Deployment`,<br/>all matched `Pod`'s egress IP in the real node is `eip` | p3 | false | | |
| R00002 | Use `kwok` to create 10 `Node`, create `Deployment` with 1000 replicas, create `Policy` and set `PodSelector` to match `Deployment`,<br/>After restarting `Deployment` successfully, all matched `Pod`'s egress IP in the real node is `eip` | p3 | false | | |
| R00005 | When the node where `eip` takes effect is shut down, `eip` will take effect to another node matching `NodeSelector`, and `egressGatewayStatus` and `EgressClusterStatus` are updated as expected, and the `EgressTunnel` corresponding to the shutdown node ` will be deleted and the egress IP will be accessed as expected | p3 | false | done | |
| R00006 | After shutting down all nodes matched by `NodeSelector` in `egressGateway`,<br/>`Pod`’s egress IP will be changed from `eip` to non-`eip`, `egressGatewayStatus.NodeList` will be empty, and the related `EgressIgnoreCIDR.NodeIP` will be deleted and the `EgressTunnel` corresponding to the shutdown node will be deleted. <br/> After one of the `node` is turned on, `egressgateway` will recover in a short time and record the recovery time, and `eip` will be revalidated as the egress IP of `Pod`, and the `nodeIP` will be added to `EgressIgnoreCIDR.NodeIP` and `node` related information in `egressGatewayStatus.NodeList` is updated correctly, <br/>after all boots, `eip` will only take effect on the first recovered `node`, and `EgressIgnoreCIDR.NodeIP` is updated correct | p3 | false | done | |
| R00007 | Restart each component in the cluster (including calico, kube-proxy) `Pod` in turn. During the restart process, the access IP to outside the cluster is the set `eip` before, and the traffic cannot be interrupted. After the cluster returns to normal, `egressgateway` The individual `cr` state of the component is correct | p1 | false | done | |
Expand Down
Loading

0 comments on commit f188111

Please sign in to comment.