Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

replication: check the up stores to switch to async #3991

Merged
merged 2 commits into from
Aug 16, 2021
Merged

replication: check the up stores to switch to async #3991

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
15e0719
replication: check the up stores to switch to async
nolouch Aug 16, 2021
914da78
Merge branch 'master' into fix-switch
ti-chi-bot Aug 16, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 24 additions & 15 deletions server/replication/replication_mode.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -369,21 +369,22 @@ func (m *ModeManager) tickDR() {

drTickCounter.Inc()

totalPrimary, totalDr := m.config.DRAutoSync.PrimaryReplicas, m.config.DRAutoSync.DRReplicas
downPrimary, downDr := m.checkStoreStatus()
totalPrimaryPeers, totalDrPeers := m.config.DRAutoSync.PrimaryReplicas, m.config.DRAutoSync.DRReplicas
downPrimaryStores, downDrStores, upPrimayStores, upDrStores := m.checkStoreStatus()

// canSync is true when every region has at least 1 replica in each DC.
canSync := downPrimary < totalPrimary && downDr < totalDr
canSync := downPrimaryStores < totalPrimaryPeers && downDrStores < totalDrPeers &&
upPrimayStores > 0 && upDrStores > 0

// hasMajority is true when every region has majority peer online.
var upPeers int
if downPrimary < totalPrimary {
upPeers += totalPrimary - downPrimary
if downPrimaryStores < totalPrimaryPeers {
upPeers += totalPrimaryPeers - downPrimaryStores
}
if downDr < totalDr {
upPeers += totalDr - downDr
if downDrStores < totalDrPeers {
upPeers += totalDrPeers - downDrStores
}
hasMajority := upPeers*2 > totalPrimary+totalDr
hasMajority := upPeers*2 > totalPrimaryPeers+totalDrPeers

// If hasMajority is false, the cluster is always unavailable. Switch to async won't help.
if !canSync && hasMajority && m.drGetState() != drStateAsync && m.drCheckAsyncTimeout() {
Expand All @@ -407,17 +408,25 @@ func (m *ModeManager) tickDR() {
}
}

func (m *ModeManager) checkStoreStatus() (primaryFailCount, drFailCount int) {
func (m *ModeManager) checkStoreStatus() (primaryDownCount, drDownCount, primaryUpCount, drUpCount int) {
m.RLock()
defer m.RUnlock()
for _, s := range m.cluster.GetStores() {
if !s.IsTombstone() && s.DownTime() >= m.config.DRAutoSync.WaitStoreTimeout.Duration {
labelValue := s.GetLabelValue(m.config.DRAutoSync.LabelKey)
if labelValue == m.config.DRAutoSync.Primary {
primaryFailCount++
down := !s.IsTombstone() && s.DownTime() >= m.config.DRAutoSync.WaitStoreTimeout.Duration
labelValue := s.GetLabelValue(m.config.DRAutoSync.LabelKey)
if labelValue == m.config.DRAutoSync.Primary {
if down {
primaryDownCount++
} else {
primaryUpCount++
}
if labelValue == m.config.DRAutoSync.DR {
drFailCount++

}
if labelValue == m.config.DRAutoSync.DR {
if down {
drDownCount++
} else {
drUpCount++
}
}
}
Expand Down
17 changes: 15 additions & 2 deletions server/replication/replication_mode_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -166,8 +166,6 @@ func (s *testReplicationMode) TestStateSwitch(c *C) {
cluster.AddLabelsStore(1, 1, map[string]string{"zone": "zone1"})
cluster.AddLabelsStore(2, 1, map[string]string{"zone": "zone1"})
cluster.AddLabelsStore(3, 1, map[string]string{"zone": "zone1"})
cluster.AddLabelsStore(4, 1, map[string]string{"zone": "zone2"})
cluster.AddLabelsStore(5, 1, map[string]string{"zone": "zone2"})

// initial state is sync
c.Assert(rep.drGetState(), Equals, drStateSync)
Expand All @@ -178,6 +176,21 @@ func (s *testReplicationMode) TestStateSwitch(c *C) {
stateID = rep.drAutoSync.StateID
}

// only one zone, sync -> async
rep.tickDR()
c.Assert(rep.drGetState(), Equals, drStateAsync)
assertStateIDUpdate()

// add new store in dr zone.
cluster.AddLabelsStore(4, 1, map[string]string{"zone": "zone2"})
cluster.AddLabelsStore(5, 1, map[string]string{"zone": "zone2"})
// async -> sync
rep.tickDR()
c.Assert(rep.drGetState(), Equals, drStateSyncRecover)
rep.drSwitchToSync()
c.Assert(rep.drGetState(), Equals, drStateSync)
assertStateIDUpdate()

// sync -> async
rep.tickDR()
c.Assert(rep.drGetState(), Equals, drStateSync)
Expand Down