diff options
| author | Paul Buetow <paul@buetow.org> | 2026-03-13 09:29:59 +0200 |
|---|---|---|
| committer | Paul Buetow <paul@buetow.org> | 2026-03-13 09:29:59 +0200 |
| commit | d8f88d455990636bb797643dee7d39a95bbbd62c (patch) | |
| tree | 8c8447fc975ec6deebe48218d27e3defa1b3dcce /internal/mapr | |
| parent | 7a79d0a8bf58b05dfbae331d00275739530b9584 (diff) | |
task 4abe7505: reset dmap generation state
Diffstat (limited to 'internal/mapr')
| -rw-r--r-- | internal/mapr/client/aggregate.go | 40 | ||||
| -rw-r--r-- | internal/mapr/client/aggregate_test.go | 70 | ||||
| -rw-r--r-- | internal/mapr/client/session_state.go | 95 | ||||
| -rw-r--r-- | internal/mapr/client/session_state_test.go | 81 |
4 files changed, 274 insertions, 12 deletions
diff --git a/internal/mapr/client/aggregate.go b/internal/mapr/client/aggregate.go index 2750643..8cbd339 100644 --- a/internal/mapr/client/aggregate.go +++ b/internal/mapr/client/aggregate.go @@ -12,30 +12,46 @@ import ( // Aggregate mapreduce data on the DTail client side. type Aggregate struct { - // This is the mapr query specified on the command line. - query *mapr.Query // This represents aggregated data of a single remote server. group *mapr.GroupSet - // This represents the merged aggregated data of all servers. - globalGroup *mapr.GlobalGroupSet + // Shared per-client session state. + session *SessionState + // The currently tracked shared generation. + generation uint64 // The server we aggregate the data for (logging and debugging purposes only) server string } // NewAggregate create new client aggregator. -func NewAggregate(server string, query *mapr.Query, - globalGroup *mapr.GlobalGroupSet) *Aggregate { +func NewAggregate(server string, session *SessionState) *Aggregate { + generation := uint64(0) + if session != nil { + generation = session.Snapshot().Generation + } return &Aggregate{ - query: query, - group: mapr.NewGroupSet(), - globalGroup: globalGroup, - server: server, + group: mapr.NewGroupSet(), + session: session, + generation: generation, + server: server, } } // Aggregate data from mapr log line into local (and global) group sets. func (a *Aggregate) Aggregate(message string) error { + if a.session == nil { + return fmt.Errorf("missing client mapreduce session state") + } + + snapshot := a.session.Snapshot() + if snapshot.Query == nil || snapshot.GlobalGroup == nil { + return fmt.Errorf("missing client mapreduce query state") + } + if snapshot.Generation != a.generation { + a.group.InitSet() + a.generation = snapshot.Generation + } + parts := strings.Split(message, protocol.AggregateDelimiter) if len(parts) < 4 { return fmt.Errorf("aggregate message without any real data") @@ -51,7 +67,7 @@ func (a *Aggregate) Aggregate(message string) error { set := a.group.GetSet(groupKey) var addedSamples bool - for _, sc := range a.query.Select { + for _, sc := range snapshot.Query.Select { if val, ok := fields[sc.FieldStorage]; ok { if err := set.Aggregate(sc.FieldStorage, sc.Operation, val, true); err != nil { dlog.Client.Error(err) @@ -65,7 +81,7 @@ func (a *Aggregate) Aggregate(message string) error { } // Merge data from group into global group. - isMerged, err := a.globalGroup.MergeNoblock(a.query, a.group) + isMerged, err := snapshot.GlobalGroup.MergeNoblock(snapshot.Query, a.group) if err != nil { return fmt.Errorf("unable to merge aggregate data for server %s: %w", a.server, err) } diff --git a/internal/mapr/client/aggregate_test.go b/internal/mapr/client/aggregate_test.go new file mode 100644 index 0000000..8ac94a1 --- /dev/null +++ b/internal/mapr/client/aggregate_test.go @@ -0,0 +1,70 @@ +package client + +import ( + "strings" + "testing" + + "github.com/mimecast/dtail/internal/mapr" + "github.com/mimecast/dtail/internal/protocol" +) + +func TestAggregateResetsPendingLocalStateOnGenerationChange(t *testing.T) { + query := mustSessionStateQuery(t, "select status,count(status) from stats group by status") + state := NewSessionState(query) + aggregate := NewAggregate("srv1", state) + countStorage := aggregateCountStorage(t, query) + + oldSet := aggregate.group.GetSet("ERROR") + oldSet.Samples = 1 + oldSet.FValues[countStorage] = 1 + + rawQuery := "select status,count(status) from warnings group by status" + if _, err := state.CommitQuery(rawQuery, 2); err != nil { + t.Fatalf("CommitQuery() error = %v", err) + } + + snapshot := state.Snapshot() + message := strings.Join([]string{ + "WARN", + "1", + aggregateCountStorage(t, snapshot.Query) + protocol.AggregateKVDelimiter + "1", + "", + }, protocol.AggregateDelimiter) + + if err := aggregate.Aggregate(message); err != nil { + t.Fatalf("Aggregate() error = %v", err) + } + + result, numRows, err := snapshot.GlobalGroup.Result(snapshot.Query, 10, nil) + if err != nil { + t.Fatalf("Result() error = %v", err) + } + if numRows != 1 { + t.Fatalf("numRows = %d, want 1", numRows) + } + if !strings.Contains(result, "1") { + t.Fatalf("expected one new-generation aggregate row, got %q", result) + } +} + +func TestAggregateRejectsMalformedMessage(t *testing.T) { + query := mustSessionStateQuery(t, "select count(status) from stats group by status") + state := NewSessionState(query) + aggregate := NewAggregate("srv1", state) + + if err := aggregate.Aggregate("broken"); err == nil { + t.Fatalf("expected Aggregate() to reject malformed messages") + } +} + +func aggregateCountStorage(t *testing.T, query *mapr.Query) string { + t.Helper() + + for _, selectCondition := range query.Select { + if selectCondition.Operation == mapr.Count { + return selectCondition.FieldStorage + } + } + t.Fatalf("query %q does not contain count() storage", query.RawQuery) + return "" +} diff --git a/internal/mapr/client/session_state.go b/internal/mapr/client/session_state.go new file mode 100644 index 0000000..1983644 --- /dev/null +++ b/internal/mapr/client/session_state.go @@ -0,0 +1,95 @@ +package client + +import ( + "fmt" + "sync" + + "github.com/mimecast/dtail/internal/mapr" +) + +// SessionSnapshot captures the current client-side mapreduce session state. +type SessionSnapshot struct { + Generation uint64 + Query *mapr.Query + GlobalGroup *mapr.GlobalGroupSet + LastResult string +} + +// SessionState keeps the mutable mapreduce query state shared by the client +// reporter and per-server handlers. +type SessionState struct { + mu sync.RWMutex + generation uint64 + query *mapr.Query + global *mapr.GlobalGroupSet + lastResult string + changedCh chan struct{} +} + +// NewSessionState returns a new shared mapreduce session state. +func NewSessionState(query *mapr.Query) *SessionState { + return &SessionState{ + query: query, + global: mapr.NewGlobalGroupSet(), + changedCh: make(chan struct{}, 1), + } +} + +// Snapshot returns a point-in-time copy of the shared mapreduce state. +func (s *SessionState) Snapshot() SessionSnapshot { + s.mu.RLock() + defer s.mu.RUnlock() + + return SessionSnapshot{ + Generation: s.generation, + Query: s.query, + GlobalGroup: s.global, + LastResult: s.lastResult, + } +} + +// Changes returns a channel that is signaled whenever a new generation is committed. +func (s *SessionState) Changes() <-chan struct{} { + return s.changedCh +} + +// CommitQuery resets the shared aggregation state for a newly accepted query generation. +func (s *SessionState) CommitQuery(rawQuery string, generation uint64) (*mapr.Query, error) { + query, err := mapr.NewQuery(rawQuery) + if err != nil { + return nil, fmt.Errorf("parse session query: %w", err) + } + + s.mu.Lock() + s.generation = generation + s.query = query + s.global = mapr.NewGlobalGroupSet() + s.lastResult = "" + s.mu.Unlock() + + s.notifyChange() + return query, nil +} + +// CommitRenderedResult stores the last rendered result for the active generation. +func (s *SessionState) CommitRenderedResult(generation uint64, result string) (changed bool, ok bool) { + s.mu.Lock() + defer s.mu.Unlock() + + if s.generation != generation { + return false, false + } + if s.lastResult == result { + return false, true + } + + s.lastResult = result + return true, true +} + +func (s *SessionState) notifyChange() { + select { + case s.changedCh <- struct{}{}: + default: + } +} diff --git a/internal/mapr/client/session_state_test.go b/internal/mapr/client/session_state_test.go new file mode 100644 index 0000000..f43ca70 --- /dev/null +++ b/internal/mapr/client/session_state_test.go @@ -0,0 +1,81 @@ +package client + +import ( + "testing" + + "github.com/mimecast/dtail/internal/mapr" +) + +func TestSessionStateCommitQueryResetsGenerationAndResults(t *testing.T) { + query := mustSessionStateQuery(t, "select count(status) from stats group by status") + state := NewSessionState(query) + + initial := state.Snapshot() + group := mapr.NewGroupSet() + set := group.GetSet("ERROR") + set.Samples = 1 + set.FValues[query.Select[0].FieldStorage] = 1 + if err := initial.GlobalGroup.Merge(query, group); err != nil { + t.Fatalf("Merge() error = %v", err) + } + if changed, ok := state.CommitRenderedResult(initial.Generation, "old-result"); !ok || !changed { + t.Fatalf("CommitRenderedResult() = changed:%v ok:%v, want changed and ok", changed, ok) + } + + rawQuery := "select count(status) from warnings group by status" + updatedQuery, err := state.CommitQuery(rawQuery, 3) + if err != nil { + t.Fatalf("CommitQuery() error = %v", err) + } + if updatedQuery == nil || updatedQuery.RawQuery != rawQuery { + t.Fatalf("unexpected updated query: %#v", updatedQuery) + } + + select { + case <-state.Changes(): + default: + t.Fatalf("expected change notification after CommitQuery") + } + + updated := state.Snapshot() + if updated.Generation != 3 { + t.Fatalf("generation = %d, want 3", updated.Generation) + } + if updated.Query == nil || updated.Query.RawQuery != rawQuery { + t.Fatalf("unexpected query after commit: %#v", updated.Query) + } + if !updated.GlobalGroup.IsEmpty() { + t.Fatalf("expected committed global group to be reset") + } + if updated.LastResult != "" { + t.Fatalf("last result = %q, want empty", updated.LastResult) + } +} + +func TestSessionStateCommitQueryRejectsInvalidQuery(t *testing.T) { + query := mustSessionStateQuery(t, "select count(status) from stats group by status") + state := NewSessionState(query) + before := state.Snapshot() + + if _, err := state.CommitQuery("select from", 5); err == nil { + t.Fatalf("expected CommitQuery() to reject invalid query") + } + + after := state.Snapshot() + if after.Generation != before.Generation { + t.Fatalf("generation changed on invalid query: got %d want %d", after.Generation, before.Generation) + } + if after.Query == nil || after.Query.RawQuery != before.Query.RawQuery { + t.Fatalf("query changed on invalid query: before=%#v after=%#v", before.Query, after.Query) + } +} + +func mustSessionStateQuery(t *testing.T, queryStr string) *mapr.Query { + t.Helper() + + query, err := mapr.NewQuery(queryStr) + if err != nil { + t.Fatalf("NewQuery(%q) error = %v", queryStr, err) + } + return query +} |
