wangjia
2080062d8c
盲区:节点 status 只看 agent 的 gRPC 在线,sing-box 数据面坏了(崩/配置坏/数据口不通) 而 agent 仍在线时,节点仍显示 up、connect 放行、客户端「已连接」但流量全失败,且不自愈。 控制面: - agent 每心跳探 sing-box clash_api /version(2s 超时,连续 2 次失败才报不健康,防抖), 经新增 HeartbeatRequest.data_plane_healthy 上报(手写 agentv1 契约 + proto 同步;JSON codec)。 - NodeLoad 加 DataPlaneHealthy(随 load 写 Redis;字段缺失默认 healthy 防滚动期误杀)。 - effectiveNodeStatus 扩为 (dbStatus, agentOnline, dataPlaneHealthy);ListNodes 与 ConnectNode 统一改用它 → 数据面坏的节点列表置灰 + connect 返回 404 拦截。 客户端(connection_provider 连通看门狗): - 连上后每 15s 经隧道 HTTP 探海外 generate_204(可注入),连续 3 次失败判当前节点不可用。 - 智能选择 → 自动切到其他最优可用节点重连;手动选定 → 断开并提示「节点异常」(尊重用户选择)。 - 新增 nodeUnhealthySwitched/nodeUnhealthyError 文案。 测试:agent 健康探测防抖、effectiveNodeStatus 真值表、dataPlaneHealthy 助手、看门狗 智能切/手动断/健康不触发;go test ./... 与 flutter test 全绿。 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
71 lines
2.3 KiB
Go
71 lines
2.3 KiB
Go
package httpapi
|
|
|
|
import (
|
|
"context"
|
|
"errors"
|
|
"testing"
|
|
|
|
"github.com/wangjia/pangolin/server/internal/nodes"
|
|
)
|
|
|
|
type fakeLoadReader struct {
|
|
load *nodes.NodeLoad
|
|
ok bool
|
|
err error
|
|
}
|
|
|
|
func (f fakeLoadReader) Get(context.Context, string) (*nodes.NodeLoad, bool, error) {
|
|
return f.load, f.ok, f.err
|
|
}
|
|
|
|
// TestDataPlaneHealthy: unknown (nil reader / missing / error) defaults to healthy
|
|
// (agent liveness gates); a present load returns its reported flag.
|
|
func TestDataPlaneHealthy(t *testing.T) {
|
|
ctx := context.Background()
|
|
cases := []struct {
|
|
name string
|
|
api *NodeAPI
|
|
want bool
|
|
}{
|
|
{"nil reader → healthy", &NodeAPI{}, true},
|
|
{"missing load → healthy", &NodeAPI{load: fakeLoadReader{ok: false}}, true},
|
|
{"read error → healthy", &NodeAPI{load: fakeLoadReader{err: errors.New("boom")}}, true},
|
|
{"present healthy", &NodeAPI{load: fakeLoadReader{load: &nodes.NodeLoad{DataPlaneHealthy: true}, ok: true}}, true},
|
|
{"present unhealthy → false", &NodeAPI{load: fakeLoadReader{load: &nodes.NodeLoad{DataPlaneHealthy: false}, ok: true}}, false},
|
|
}
|
|
for _, c := range cases {
|
|
t.Run(c.name, func(t *testing.T) {
|
|
if got := c.api.dataPlaneHealthy(ctx, "n1"); got != c.want {
|
|
t.Errorf("dataPlaneHealthy = %v, want %v", got, c.want)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
// TestEffectiveNodeStatus guards the "don't show a node healthy when it can't
|
|
// serve" fix: a DB-'up' node reports "down" when its agent is offline OR its data
|
|
// plane (sing-box) is unhealthy.
|
|
func TestEffectiveNodeStatus(t *testing.T) {
|
|
cases := []struct {
|
|
name string
|
|
db string
|
|
online bool
|
|
healthy bool
|
|
want string
|
|
}{
|
|
{"up + agent online + healthy", "up", true, true, "up"},
|
|
{"up + agent offline → down", "up", false, true, "down"}, // agent-liveness fix
|
|
{"up + agent online + dp unhealthy → down", "up", true, false, "down"}, // data-plane fix
|
|
{"up + offline + unhealthy → down", "up", false, false, "down"},
|
|
{"draining untouched", "draining", false, false, "draining"},
|
|
{"down stays down", "down", true, true, "down"},
|
|
}
|
|
for _, c := range cases {
|
|
t.Run(c.name, func(t *testing.T) {
|
|
if got := effectiveNodeStatus(c.db, c.online, c.healthy); got != c.want {
|
|
t.Errorf("effectiveNodeStatus(%q, %v, %v) = %q, want %q", c.db, c.online, c.healthy, got, c.want)
|
|
}
|
|
})
|
|
}
|
|
}
|