fix(node): give each node one push budget per traffic sync

After a sync tick disabled clients, the master pushed every affected
node inbound with its own 4 s timeout, a few at a time, and the job
served no other node until all were done; a hanging node could stall it
for as long as it had inbounds to report. Pushes are now grouped by node
and each node shares one budget, still a few inbounds at a time; what
does not make it stays for the reconcile, the node being dirty already.
This commit is contained in:
MHSanaei
2026-10-31 14:07:08 +01:00
parent d005d89a8b
commit 4d27313c5e
2 changed files with 59 additions and 13 deletions
+29 -13
View File
@@ -2,6 +2,7 @@ package service
import (
"context"
"slices"
"strings"
"github.com/mhsanaei/3x-ui/v3/internal/database/model"
@@ -62,21 +63,36 @@ func (b *trafficMutationBatch) markNodesTx(tx *gorm.DB) error {
// applyTrafficRemotePlans is bounded like every per-client node push: the nodes are
// already dirty, so an offline or slow one defers to the reconcile.
func (s *InboundService) applyTrafficRemotePlans(plans []trafficInboundUpdatePlan) bool {
ids := make([]int, len(plans))
byNode := make(map[int][]int)
var nodeIDs []int
for i := range plans {
ids[i] = plans[i].newInbound.Id
nodeID := 0
if plans[i].newInbound.NodeID != nil {
nodeID = *plans[i].newInbound.NodeID
}
if _, seen := byNode[nodeID]; !seen {
nodeIDs = append(nodeIDs, nodeID)
}
byNode[nodeID] = append(byNode[nodeID], i)
}
failed, panics := fanoutInboundResults(ids, inboundFanoutConcurrency, func(i int) bool {
rt, push, _, err := s.nodePushPlan(&plans[i].newInbound)
if err == nil && push {
ctx, cancel := nodePushContext()
err = rt.UpdateInbound(ctx, &plans[i].oldInbound, &plans[i].newInbound)
cancel()
}
if err != nil {
logger.Debug("traffic post-commit remote apply failed:", err)
}
return err != nil
failed, panics := fanoutInboundResults(nodeIDs, nodeFanoutConcurrency, func(n int) bool {
// One budget per node, not per inbound: a node choosing how many inbounds it reports
// must not choose how long the sync job waits on it before serving other nodes.
ctx, cancel := nodePushContext()
defer cancel()
indexes := byNode[nodeIDs[n]]
nodeFailed, nodePanics := fanoutInboundResults(indexes, inboundFanoutConcurrency, func(k int) bool {
plan := &plans[indexes[k]]
rt, push, _, err := s.nodePushPlan(&plan.newInbound)
if err == nil && push {
err = rt.UpdateInbound(ctx, &plan.oldInbound, &plan.newInbound)
}
if err != nil {
logger.Debug("traffic post-commit remote apply failed:", err)
}
return err != nil
})
return slices.Contains(nodeFailed, true) || slices.ContainsFunc(nodePanics, func(p error) bool { return p != nil })
})
needRestart := false
for i := range failed {