wip - remove retry logic

Signed-off-by: Florent Poinsard <[email protected]>
vitessio · Jan 21, 2025 · 2b75327 · 2b75327
1 parent 93d6c52
commit 2b75327
Show file tree

Hide file tree

Showing 2 changed files with 48 additions and 48 deletions.
diff --git a/go/cmd/vtbackup/cli/vtbackup.go b/go/cmd/vtbackup/cli/vtbackup.go
@@ -537,12 +537,12 @@ func takeBackup(ctx, backgroundCtx context.Context, topoServer *topo.Server, bac
 
 		waitStartTime = time.Now()
 
-		continuousErrorCount int
+		// continuousErrorCount int
 	)
 	for {
-		if continuousErrorCount == maximumErrorCountWhenWaitingForReplicationStatus {
-			return fmt.Errorf("timeout waiting for replication status after %d errors", maximumErrorCountWhenWaitingForReplicationStatus)
-		}
+		// if continuousErrorCount == maximumErrorCountWhenWaitingForReplicationStatus {
+		// 	return fmt.Errorf("timeout waiting for replication status after %d errors", maximumErrorCountWhenWaitingForReplicationStatus)
+		// }
 
 		select {
 		case <-ctx.Done():
@@ -554,7 +554,7 @@ func takeBackup(ctx, backgroundCtx context.Context, topoServer *topo.Server, bac
 		status, statusErr = mysqld.ReplicationStatus(ctx)
 		if statusErr != nil {
 			log.Warningf("Error getting replication status: %v", statusErr)
-			continuousErrorCount++
+			// continuousErrorCount++
 			continue
 		}
 		if status.Position.AtLeast(primaryPos) {
@@ -577,11 +577,11 @@ func takeBackup(ctx, backgroundCtx context.Context, topoServer *topo.Server, bac
 			if err := startReplication(ctx, mysqld, topoServer); err != nil {
 				log.Warningf("Failed to restart replication: %v", err)
 			}
-			continuousErrorCount++
+			// continuousErrorCount++
 		} else {
 			// Since replication is working if we got here, let's reset the error count to zero.
 			// This allows us to avoid failing if we only have transient errors from time to time.
-			continuousErrorCount = 0
+			// continuousErrorCount = 0
 			phaseStatus.Set([]string{phaseNameCatchupReplication, phaseStatusCatchupReplicationStopped}, 0)
 		}
 	}

diff --git a/go/test/endtoend/backup/vtbackup/backup_only_test.go b/go/test/endtoend/backup/vtbackup/backup_only_test.go
@@ -47,47 +47,47 @@ var (
 		) Engine=InnoDB;`
 )
 
-func TestFailingReplication(t *testing.T) {
-	prepareCluster(t)
-
-	// Run the entire backup test
-	firstBackupTest(t, false)
-
-	// Insert one more row, the primary will be ahead of the last backup
-	_, err := primary.VttabletProcess.QueryTablet("insert into vt_insert_test (msg) values ('test_failure')", keyspaceName, true)
-	require.NoError(t, err)
-
-	// Disable replication from the primary by removing the grants to 'vt_repl'.
-	_, err = primary.VttabletProcess.QueryTablet("REVOKE REPLICATION SLAVE ON *.* FROM 'vt_repl'@'%';", keyspaceName, true)
-	require.NoError(t, err)
-	_, err = primary.VttabletProcess.QueryTablet("FLUSH PRIVILEGES;", keyspaceName, true)
-	require.NoError(t, err)
-
-	// Take a backup with vtbackup: the process should fail entirely as it cannot replicate from the primary.
-	_, err = startVtBackup(t, false, false, false)
-	require.Error(t, err)
-
-	// keep in mind how many backups we have right now
-	backups, err := listBackups(shardKsName)
-	require.NoError(t, err)
-
-	// In 30 seconds, grant the replication permission again to 'vt_repl'.
-	// This will mean that vtbackup should fail to replicate for ~30 seconds, until we grant the permission again.
-	go func() {
-		<-time.After(30 * time.Second)
-		_, err = primary.VttabletProcess.QueryTablet("GRANT REPLICATION SLAVE ON *.* TO 'vt_repl'@'%';", keyspaceName, true)
-		require.NoError(t, err)
-		_, err = primary.VttabletProcess.QueryTablet("FLUSH PRIVILEGES;", keyspaceName, true)
-		require.NoError(t, err)
-	}()
-
-	// this will initially be stuck trying to replicate from the primary, and once we re-grant the permission in
-	// the goroutine above, the process will work and complete successfully.
-	_ = vtBackup(t, false, false, false)
-	verifyBackupCount(t, shardKsName, len(backups)+1)
-
-	tearDown(t, true)
-}
+// func TestFailingReplication(t *testing.T) {
+// 	prepareCluster(t)
+//
+// 	// Run the entire backup test
+// 	firstBackupTest(t, false)
+//
+// 	// Insert one more row, the primary will be ahead of the last backup
+// 	_, err := primary.VttabletProcess.QueryTablet("insert into vt_insert_test (msg) values ('test_failure')", keyspaceName, true)
+// 	require.NoError(t, err)
+//
+// 	// Disable replication from the primary by removing the grants to 'vt_repl'.
+// 	_, err = primary.VttabletProcess.QueryTablet("REVOKE REPLICATION SLAVE ON *.* FROM 'vt_repl'@'%';", keyspaceName, true)
+// 	require.NoError(t, err)
+// 	_, err = primary.VttabletProcess.QueryTablet("FLUSH PRIVILEGES;", keyspaceName, true)
+// 	require.NoError(t, err)
+//
+// 	// Take a backup with vtbackup: the process should fail entirely as it cannot replicate from the primary.
+// 	_, err = startVtBackup(t, false, false, false)
+// 	require.Error(t, err)
+//
+// 	// keep in mind how many backups we have right now
+// 	backups, err := listBackups(shardKsName)
+// 	require.NoError(t, err)
+//
+// 	// In 30 seconds, grant the replication permission again to 'vt_repl'.
+// 	// This will mean that vtbackup should fail to replicate for ~30 seconds, until we grant the permission again.
+// 	go func() {
+// 		<-time.After(30 * time.Second)
+// 		_, err = primary.VttabletProcess.QueryTablet("GRANT REPLICATION SLAVE ON *.* TO 'vt_repl'@'%';", keyspaceName, true)
+// 		require.NoError(t, err)
+// 		_, err = primary.VttabletProcess.QueryTablet("FLUSH PRIVILEGES;", keyspaceName, true)
+// 		require.NoError(t, err)
+// 	}()
+//
+// 	// this will initially be stuck trying to replicate from the primary, and once we re-grant the permission in
+// 	// the goroutine above, the process will work and complete successfully.
+// 	_ = vtBackup(t, false, false, false)
+// 	verifyBackupCount(t, shardKsName, len(backups)+1)
+//
+// 	tearDown(t, true)
+// }
 
 func TestTabletInitialBackup(t *testing.T) {
 	// Test Initial Backup Flow