slack-21.0: backport v22 VTOrc fixes, part 1 (#610)

* Move to native sqlite3 queries (vitessio#17124) Signed-off-by: Dirkjan Bussink <d.bussink@gmail.com> Co-authored-by: Tim Vaillancourt <tim@timvaillancourt.com> * Improve efficiency of `vtorc` topo calls (vitessio#17071) Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> Co-authored-by: Matt Lord <mattalord@gmail.com> * Ensure all topo read calls consider `--topo_read_concurrency` (vitessio#17276) Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * Avoid flaky topo concurrency test (vitessio#17407) Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * `vtorc`: fetch all tablets from cells once + filter during refresh (vitessio#17388) Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * Support KeyRange in `--clusters_to_watch` flag (vitessio#17604) Signed-off-by: Manan Gupta <manan@planetscale.com> * `vtorc`: improve handling of partial cell topo results (vitessio#17718) Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * Add stats for shards watched by VTOrc Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * add more tests Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * cleanup Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * fix ineffassign Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * fix test for v21 Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> * Use prefix in all vtorc check and recover logs (vitessio#17526) Signed-off-by: Eduardo J. Ortega U. <5791035+ejortegau@users.noreply.github.com> --------- Signed-off-by: Dirkjan Bussink <d.bussink@gmail.com> Signed-off-by: Tim Vaillancourt <tim@timvaillancourt.com> Signed-off-by: Manan Gupta <manan@planetscale.com> Signed-off-by: Eduardo J. Ortega U. <5791035+ejortegau@users.noreply.github.com> Co-authored-by: Dirkjan Bussink <d.bussink@gmail.com> Co-authored-by: Matt Lord <mattalord@gmail.com> Co-authored-by: Manan Gupta <35839558+GuptaManan100@users.noreply.github.com> Co-authored-by: Eduardo J. Ortega U. <5791035+ejortegau@users.noreply.github.com>
slackhq · Feb 20, 2025 · 1f5ac88 · 1f5ac88
1 parent cf09832
commit 1f5ac88
Show file tree

Hide file tree

Showing 40 changed files with 1,491 additions and 1,026 deletions.
diff --git a/go/flags/endtoend/vtbackup.txt b/go/flags/endtoend/vtbackup.txt
@@ -231,6 +231,7 @@ Flags:
       --topo_global_root string                                     the path of the global topology data in the global topology server
       --topo_global_server_address string                           the address of the global topology server
       --topo_implementation string                                  the topology implementation to use
+      --topo_read_concurrency int                                   Maximum concurrency of topo reads per global or local cell. (default 32)
       --topo_zk_auth_file string                                    auth to use when connecting to the zk topo server, file contents should be <scheme>:<auth>, e.g., digest:user:pass
       --topo_zk_base_timeout duration                               zk base timeout (see zk.Connect) (default 30s)
       --topo_zk_max_concurrency int                                 maximum number of pending requests to send to a Zookeeper server. (default 64)

diff --git a/go/flags/endtoend/vtcombo.txt b/go/flags/endtoend/vtcombo.txt
@@ -374,7 +374,7 @@ Flags:
       --topo_global_root string                                          the path of the global topology data in the global topology server
       --topo_global_server_address string                                the address of the global topology server
       --topo_implementation string                                       the topology implementation to use
-      --topo_read_concurrency int                                        Concurrency of topo reads. (default 32)
+      --topo_read_concurrency int                                        Maximum concurrency of topo reads per global or local cell. (default 32)
       --topo_zk_auth_file string                                         auth to use when connecting to the zk topo server, file contents should be <scheme>:<auth>, e.g., digest:user:pass
       --topo_zk_base_timeout duration                                    zk base timeout (see zk.Connect) (default 30s)
       --topo_zk_max_concurrency int                                      maximum number of pending requests to send to a Zookeeper server. (default 64)

diff --git a/go/flags/endtoend/vtctld.txt b/go/flags/endtoend/vtctld.txt
@@ -164,7 +164,7 @@ Flags:
       --topo_global_root string                                          the path of the global topology data in the global topology server
       --topo_global_server_address string                                the address of the global topology server
       --topo_implementation string                                       the topology implementation to use
-      --topo_read_concurrency int                                        Concurrency of topo reads. (default 32)
+      --topo_read_concurrency int                                        Maximum concurrency of topo reads per global or local cell. (default 32)
       --topo_zk_auth_file string                                         auth to use when connecting to the zk topo server, file contents should be <scheme>:<auth>, e.g., digest:user:pass
       --topo_zk_base_timeout duration                                    zk base timeout (see zk.Connect) (default 30s)
       --topo_zk_max_concurrency int                                      maximum number of pending requests to send to a Zookeeper server. (default 64)

diff --git a/go/flags/endtoend/vtgate.txt b/go/flags/endtoend/vtgate.txt
@@ -223,7 +223,7 @@ Flags:
       --topo_global_root string                                          the path of the global topology data in the global topology server
       --topo_global_server_address string                                the address of the global topology server
       --topo_implementation string                                       the topology implementation to use
-      --topo_read_concurrency int                                        Concurrency of topo reads. (default 32)
+      --topo_read_concurrency int                                        Maximum concurrency of topo reads per global or local cell. (default 32)
       --topo_zk_auth_file string                                         auth to use when connecting to the zk topo server, file contents should be <scheme>:<auth>, e.g., digest:user:pass
       --topo_zk_base_timeout duration                                    zk base timeout (see zk.Connect) (default 30s)
       --topo_zk_max_concurrency int                                      maximum number of pending requests to send to a Zookeeper server. (default 64)

diff --git a/go/flags/endtoend/vtorc.txt b/go/flags/endtoend/vtorc.txt
@@ -24,7 +24,7 @@ Flags:
       --bind-address string                                         Bind address for the server. If empty, the server will listen on all available unicast and anycast IP addresses of the local system.
       --catch-sigpipe                                               catch and ignore SIGPIPE on stdout and stderr if specified
       --change-tablets-with-errant-gtid-to-drained                  Whether VTOrc should be changing the type of tablets with errant GTIDs to DRAINED
-      --clusters_to_watch strings                                   Comma-separated list of keyspaces or keyspace/shards that this instance will monitor and repair. Defaults to all clusters in the topology. Example: "ks1,ks2/-80"
+      --clusters_to_watch strings                                   Comma-separated list of keyspaces or keyspace/keyranges that this instance will monitor and repair. Defaults to all clusters in the topology. Example: "ks1,ks2/-80"
       --config string                                               config file name
       --config-file string                                          Full path of the config file (with extension) to use. If set, --config-path, --config-type, and --config-name are ignored.
       --config-file-not-found-handling ConfigFileNotFoundHandling   Behavior when a config file is not found. (Options: error, exit, ignore, warn) (default warn)
@@ -99,6 +99,7 @@ Flags:
       --topo_global_root string                                     the path of the global topology data in the global topology server
       --topo_global_server_address string                           the address of the global topology server
       --topo_implementation string                                  the topology implementation to use
+      --topo_read_concurrency int                                   Maximum concurrency of topo reads per global or local cell. (default 32)
       --topo_zk_auth_file string                                    auth to use when connecting to the zk topo server, file contents should be <scheme>:<auth>, e.g., digest:user:pass
       --topo_zk_base_timeout duration                               zk base timeout (see zk.Connect) (default 30s)
       --topo_zk_max_concurrency int                                 maximum number of pending requests to send to a Zookeeper server. (default 64)

diff --git a/go/flags/endtoend/vttablet.txt b/go/flags/endtoend/vttablet.txt
@@ -376,6 +376,7 @@ Flags:
       --topo_global_root string                                          the path of the global topology data in the global topology server
       --topo_global_server_address string                                the address of the global topology server
       --topo_implementation string                                       the topology implementation to use
+      --topo_read_concurrency int                                        Maximum concurrency of topo reads per global or local cell. (default 32)
       --topo_zk_auth_file string                                         auth to use when connecting to the zk topo server, file contents should be <scheme>:<auth>, e.g., digest:user:pass
       --topo_zk_base_timeout duration                                    zk base timeout (see zk.Connect) (default 30s)
       --topo_zk_max_concurrency int                                      maximum number of pending requests to send to a Zookeeper server. (default 64)

diff --git a/go/vt/discovery/healthcheck.go b/go/vt/discovery/healthcheck.go
@@ -376,7 +376,7 @@ func NewHealthCheck(ctx context.Context, retryDelay, healthCheckTimeout time.Dur
 		if c == "" {
 			continue
 		}
-		topoWatchers = append(topoWatchers, NewTopologyWatcher(ctx, topoServer, hc, filters, c, refreshInterval, refreshKnownTablets, topo.DefaultConcurrency))
+		topoWatchers = append(topoWatchers, NewTopologyWatcher(ctx, topoServer, hc, filters, c, refreshInterval, refreshKnownTablets))
 	}
 
 	hc.topoWatchers = topoWatchers

diff --git a/go/vt/discovery/topology_watcher.go b/go/vt/discovery/topology_watcher.go
@@ -26,16 +26,13 @@ import (
 	"sync"
 	"time"
 
-	"vitess.io/vitess/go/vt/topo/topoproto"
-
-	"vitess.io/vitess/go/vt/key"
-
 	"vitess.io/vitess/go/stats"
 	"vitess.io/vitess/go/trace"
-
+	"vitess.io/vitess/go/vt/key"
 	"vitess.io/vitess/go/vt/log"
-	"vitess.io/vitess/go/vt/proto/topodata"
+	topodatapb "vitess.io/vitess/go/vt/proto/topodata"
 	"vitess.io/vitess/go/vt/topo"
+	"vitess.io/vitess/go/vt/topo/topoproto"
 )
 
 const (
@@ -56,7 +53,7 @@ var (
 // tabletInfo is used internally by the TopologyWatcher struct.
 type tabletInfo struct {
 	alias  string
-	tablet *topodata.Tablet
+	tablet *topodatapb.Tablet
 }
 
 // TopologyWatcher polls the topology periodically for changes to
@@ -70,7 +67,6 @@ type TopologyWatcher struct {
 	cell                string
 	refreshInterval     time.Duration
 	refreshKnownTablets bool
-	concurrency         int
 	ctx                 context.Context
 	cancelFunc          context.CancelFunc
 	// wg keeps track of all launched Go routines.
@@ -92,15 +88,14 @@ type TopologyWatcher struct {
 
 // NewTopologyWatcher returns a TopologyWatcher that monitors all
 // the tablets in a cell, and reloads them as needed.
-func NewTopologyWatcher(ctx context.Context, topoServer *topo.Server, hc HealthCheck, f TabletFilter, cell string, refreshInterval time.Duration, refreshKnownTablets bool, topoReadConcurrency int) *TopologyWatcher {
+func NewTopologyWatcher(ctx context.Context, topoServer *topo.Server, hc HealthCheck, f TabletFilter, cell string, refreshInterval time.Duration, refreshKnownTablets bool) *TopologyWatcher {
 	tw := &TopologyWatcher{
 		topoServer:          topoServer,
 		healthcheck:         hc,
 		tabletFilter:        f,
 		cell:                cell,
 		refreshInterval:     refreshInterval,
 		refreshKnownTablets: refreshKnownTablets,
-		concurrency:         topoReadConcurrency,
 		tablets:             make(map[string]*tabletInfo),
 	}
 	tw.firstLoadChan = make(chan struct{})
@@ -112,7 +107,7 @@ func NewTopologyWatcher(ctx context.Context, topoServer *topo.Server, hc HealthC
 }
 
 func (tw *TopologyWatcher) getTablets() ([]*topo.TabletInfo, error) {
-	return tw.topoServer.GetTabletsByCell(tw.ctx, tw.cell, &topo.GetTabletsByCellOptions{Concurrency: tw.concurrency})
+	return tw.topoServer.GetTabletsByCell(tw.ctx, tw.cell, nil)
 }
 
 // Start starts the topology watcher.
@@ -271,14 +266,14 @@ func (tw *TopologyWatcher) TopoChecksum() uint32 {
 // to be applied as an additional filter on the list of tablets returned by its getTablets function.
 type TabletFilter interface {
 	// IsIncluded returns whether tablet is included in this filter
-	IsIncluded(tablet *topodata.Tablet) bool
+	IsIncluded(tablet *topodatapb.Tablet) bool
 }
 
 // TabletFilters contains filters for tablets.
 type TabletFilters []TabletFilter
 
 // IsIncluded returns true if a tablet passes all filters.
-func (tf TabletFilters) IsIncluded(tablet *topodata.Tablet) bool {
+func (tf TabletFilters) IsIncluded(tablet *topodatapb.Tablet) bool {
 	for _, filter := range tf {
 		if !filter.IsIncluded(tablet) {
 			return false
@@ -299,7 +294,7 @@ type FilterByShard struct {
 type filterShard struct {
 	keyspace string
 	shard    string
-	keyRange *topodata.KeyRange // only set if shard is also a KeyRange
+	keyRange *topodatapb.KeyRange // only set if shard is also a KeyRange
 }
 
 // NewFilterByShard creates a new FilterByShard for use by a
@@ -344,7 +339,7 @@ func NewFilterByShard(filters []string) (*FilterByShard, error) {
 }
 
 // IsIncluded returns true iff the tablet's keyspace and shard match what we have.
-func (fbs *FilterByShard) IsIncluded(tablet *topodata.Tablet) bool {
+func (fbs *FilterByShard) IsIncluded(tablet *topodatapb.Tablet) bool {
 	canonical, kr, err := topo.ValidateShardName(tablet.Shard)
 	if err != nil {
 		log.Errorf("Error parsing shard name %v, will ignore tablet: %v", tablet.Shard, err)
@@ -384,7 +379,7 @@ func NewFilterByKeyspace(selectedKeyspaces []string) *FilterByKeyspace {
 }
 
 // IsIncluded returns true if the tablet's keyspace matches what we have.
-func (fbk *FilterByKeyspace) IsIncluded(tablet *topodata.Tablet) bool {
+func (fbk *FilterByKeyspace) IsIncluded(tablet *topodatapb.Tablet) bool {
 	_, exist := fbk.keyspaces[tablet.Keyspace]
 	return exist
 }
@@ -403,7 +398,7 @@ func NewFilterByTabletTags(tabletTags map[string]string) *FilterByTabletTags {
 }
 
 // IsIncluded returns true if the tablet's tags match what we expect.
-func (fbtg *FilterByTabletTags) IsIncluded(tablet *topodata.Tablet) bool {
+func (fbtg *FilterByTabletTags) IsIncluded(tablet *topodatapb.Tablet) bool {
 	if fbtg.tags == nil {
 		return true
 	}

diff --git a/go/vt/discovery/topology_watcher_test.go b/go/vt/discovery/topology_watcher_test.go
@@ -67,7 +67,7 @@ func TestStartAndCloseTopoWatcher(t *testing.T) {
 	fhc := NewFakeHealthCheck(nil)
 	defer fhc.Close()
 	topologyWatcherOperations.ZeroAll()
-	tw := NewTopologyWatcher(context.Background(), ts, fhc, nil, "aa", 100*time.Microsecond, true, 5)
+	tw := NewTopologyWatcher(context.Background(), ts, fhc, nil, "aa", 100*time.Microsecond, true)
 
 	done := make(chan bool, 3)
 	result := make(chan bool, 1)
@@ -127,7 +127,7 @@ func checkWatcher(t *testing.T, refreshKnownTablets bool) {
 	logger := logutil.NewMemoryLogger()
 	topologyWatcherOperations.ZeroAll()
 	counts := topologyWatcherOperations.Counts()
-	tw := NewTopologyWatcher(context.Background(), ts, fhc, filter, "aa", 10*time.Minute, refreshKnownTablets, 5)
+	tw := NewTopologyWatcher(context.Background(), ts, fhc, filter, "aa", 10*time.Minute, refreshKnownTablets)
 
 	counts = checkOpCounts(t, counts, map[string]int64{})
 	checkChecksum(t, tw, 0)
@@ -421,7 +421,7 @@ func TestFilterByKeyspace(t *testing.T) {
 	f := TabletFilters{NewFilterByKeyspace(testKeyspacesToWatch)}
 	ts := memorytopo.NewServer(ctx, testCell)
 	defer ts.Close()
-	tw := NewTopologyWatcher(context.Background(), ts, hc, f, testCell, 10*time.Minute, true, 5)
+	tw := NewTopologyWatcher(context.Background(), ts, hc, f, testCell, 10*time.Minute, true)
 
 	for _, test := range testFilterByKeyspace {
 		// Add a new tablet to the topology.
@@ -502,7 +502,7 @@ func TestFilterByKeyspaceSkipsIgnoredTablets(t *testing.T) {
 	topologyWatcherOperations.ZeroAll()
 	counts := topologyWatcherOperations.Counts()
 	f := TabletFilters{NewFilterByKeyspace(testKeyspacesToWatch)}
-	tw := NewTopologyWatcher(context.Background(), ts, fhc, f, "aa", 10*time.Minute, false /*refreshKnownTablets*/, 5)
+	tw := NewTopologyWatcher(context.Background(), ts, fhc, f, "aa", 10*time.Minute, false /*refreshKnownTablets*/)
 
 	counts = checkOpCounts(t, counts, map[string]int64{})
 	checkChecksum(t, tw, 0)
@@ -639,7 +639,7 @@ func TestGetTabletErrorDoesNotRemoveFromHealthcheck(t *testing.T) {
 	defer fhc.Close()
 	topologyWatcherOperations.ZeroAll()
 	counts := topologyWatcherOperations.Counts()
-	tw := NewTopologyWatcher(context.Background(), ts, fhc, nil, "aa", 10*time.Minute, true, 5)
+	tw := NewTopologyWatcher(context.Background(), ts, fhc, nil, "aa", 10*time.Minute, true)
 	defer tw.Stop()
 
 	// Force fallback to getting tablets individually.

diff --git a/go/vt/external/golib/sqlutils/dialect.go b/go/vt/external/golib/sqlutils/dialect.go