Fix(increment): Fix readTs less than minTs (#6317)

Use the readTs for reads via LocalCache. This would ensure that users don't get a lot of ReadTs less than MinTs errors. This PR also fixes the flaky test in increment_test.go.

This has a side-effect on posting list cache. That cache would need to understand versioning to be effective.

Co-authored-by: Daniel Mai <daniel@dgraph.io>
(cherry picked from commit 2de4675)

dgraph/cmd/counter/increment.go → dgraph/cmd/increment/increment.go

@@ -14,10 +14,10 @@
  * limitations under the License.
  */
 
-// Package counter builds a tool that retrieves a value for UID=0x01, and increments
+// Package increment builds a tool that retrieves a value for UID=0x01, and increments
 // it by 1. If successful, it prints out the incremented value. It assumes that it has
 // access to UID=0x01, and that `val` predicate is of type int.
-package counter
+package increment
 
 import (
 	"context"

dgraph/cmd/counter/increment_test.go → dgraph/cmd/increment/increment_test.go

@@ -14,7 +14,7 @@
  * limitations under the License.
  */
 
-package counter
+package increment
 
 import (
 	"context"

dgraph/cmd/root.go

@@ -25,9 +25,9 @@ import (
 	"github.com/dgraph-io/dgraph/dgraph/cmd/bulk"
 	"github.com/dgraph-io/dgraph/dgraph/cmd/cert"
 	"github.com/dgraph-io/dgraph/dgraph/cmd/conv"
-	"github.com/dgraph-io/dgraph/dgraph/cmd/counter"
 	"github.com/dgraph-io/dgraph/dgraph/cmd/debug"
 	"github.com/dgraph-io/dgraph/dgraph/cmd/debuginfo"
+	"github.com/dgraph-io/dgraph/dgraph/cmd/increment"
 	"github.com/dgraph-io/dgraph/dgraph/cmd/live"
 	"github.com/dgraph-io/dgraph/dgraph/cmd/migrate"
 	"github.com/dgraph-io/dgraph/dgraph/cmd/version"
@@ -75,7 +75,7 @@ var rootConf = viper.New()
 // subcommands initially contains all default sub-commands.
 var subcommands = []*x.SubCommand{
 	&bulk.Bulk, &cert.Cert, &conv.Conv, &live.Live, &alpha.Alpha, &zero.Zero, &version.Version,
-	&debug.Debug, &counter.Increment, &migrate.Migrate, &debuginfo.DebugInfo, &upgrade.Upgrade,
+	&debug.Debug, &increment.Increment, &migrate.Migrate, &debuginfo.DebugInfo, &upgrade.Upgrade,
 }
 
 func initCmds() {

posting/lists.go

@@ -20,7 +20,6 @@ import (
 	"context"
 	"fmt"
 	"io/ioutil"
-	"math"
 	"os"
 	"os/exec"
 	"runtime"
@@ -181,6 +180,12 @@ func NewLocalCache(startTs uint64) *LocalCache {
 	}
 }
 
+// NoCache returns a new LocalCache instance, which won't cache anything. Useful to pass startTs
+// around.
+func NoCache(startTs uint64) *LocalCache {
+	return &LocalCache{startTs: startTs}
+}
+
 func (lc *LocalCache) getNoStore(key string) *List {
 	lc.RLock()
 	defer lc.RUnlock()
@@ -205,8 +210,8 @@ func (lc *LocalCache) SetIfAbsent(key string, updated *List) *List {
 }
 
 func (lc *LocalCache) getInternal(key []byte, readFromDisk bool) (*List, error) {
-	if lc == nil {
-		return getNew(key, pstore, math.MaxUint64)
+	if lc.plists == nil {
+		return getNew(key, pstore, lc.startTs)
 	}
 	skey := string(key)
 	if pl := lc.getNoStore(skey); pl != nil {

worker/task.go

@@ -881,6 +881,9 @@ func processTask(ctx context.Context, q *pb.Query, gid uint32) (*pb.Result, erro
 	if q.Cache == UseTxnCache {
 		qs.cache = posting.Oracle().CacheAt(q.ReadTs)
 	}
+	if qs.cache == nil {
+		qs.cache = posting.NoCache(q.ReadTs)
+	}
 	// For now, remove the query level cache. It is causing contention for queries with high
 	// fan-out.