jobs,migration: introduce long running migration job

This commit introduces a job to run long-running migration. This empowers
long-running migrations with leases, pausability, and cancelability.

Fixes #58183

Release note: None

pkg/jobs/adopt.go

@@ -25,13 +25,22 @@ import (
 	"github.com/cockroachdb/errors"
 )
 
-const claimableStatusTupleString = `(` +
-	`'` + string(StatusRunning) + `', ` +
-	`'` + string(StatusPending) + `', ` +
-	`'` + string(StatusCancelRequested) + `', ` +
-	`'` + string(StatusPauseRequested) + `', ` +
-	`'` + string(StatusReverting) + `'` +
-	`)`
+const (
+	claimableStatusList = `'` + string(StatusRunning) + `', ` +
+		`'` + string(StatusPending) + `', ` +
+		`'` + string(StatusCancelRequested) + `', ` +
+		`'` + string(StatusPauseRequested) + `', ` +
+		`'` + string(StatusReverting) + `'`
+
+	claimableStatusTupleString = `(` + claimableStatusList + `)`
+
+	nonTerminalStatusList = claimableStatusList + `, ` +
+		`'` + string(StatusPaused) + `'`
+
+	// NonTerminalStatusTupleString is a sql tuple corresponding to statuses of
+	// non-terminal jobs.
+	NonTerminalStatusTupleString = `(` + nonTerminalStatusList + `)`
+)
 
 // claimJobs places a claim with the given SessionID to job rows that are
 // available.

pkg/jobs/jobspb/BUILD.bazel

@@ -24,6 +24,7 @@ proto_library(
     strip_import_prefix = "/pkg",
     visibility = ["//visibility:public"],
     deps = [
+        "//pkg/clusterversion:clusterversion_proto",
         "//pkg/roachpb:roachpb_proto",
         "//pkg/sql/catalog/descpb:descpb_proto",
         "//pkg/sql/schemachanger/scpb:scpb_proto",
@@ -42,6 +43,7 @@ go_proto_library(
     visibility = ["//visibility:public"],
     deps = [
         "//pkg/ccl/streamingccl",  # keep
+        "//pkg/clusterversion",
         "//pkg/roachpb",
         "//pkg/security",  # keep
         "//pkg/sql/catalog/descpb",

pkg/jobs/jobspb/jobs.proto

@@ -20,6 +20,7 @@ import "sql/catalog/descpb/structured.proto";
 import "sql/catalog/descpb/tenant.proto";
 import "util/hlc/timestamp.proto";
 import "sql/schemachanger/scpb/scpb.proto";
+import "clusterversion/cluster_version.proto";
 
 message Lease {
   option (gogoproto.equal) = true;
@@ -322,6 +323,7 @@ message NewSchemaChangeProgress {
   repeated cockroach.sql.schemachanger.scpb.State states = 1;
 }
 
+
 message ResumeSpanList {
   repeated roachpb.Span resume_spans = 1 [(gogoproto.nullable) = false];
 }
@@ -577,6 +579,14 @@ message CreateStatsProgress {
 
 }
 
+message MigrationDetails {
+  clusterversion.ClusterVersion cluster_version = 1;
+}
+
+message MigrationProgress {
+
+}
+
 message Payload {
   string description = 1;
   // If empty, the description is assumed to be the statement.
@@ -619,6 +629,7 @@ message Payload {
     TypeSchemaChangeDetails typeSchemaChange = 22;
     StreamIngestionDetails streamIngestion = 23;
     NewSchemaChangeDetails newSchemaChange = 24;
+    MigrationDetails migration = 25;
   }
 }
 
@@ -641,6 +652,7 @@ message Progress {
     TypeSchemaChangeProgress typeSchemaChange = 17;
     StreamIngestionProgress streamIngest = 18;
     NewSchemaChangeProgress newSchemaChange = 19;
+    MigrationProgress migration = 20;
   }
 }
 
@@ -662,6 +674,7 @@ enum Type {
   TYPEDESC_SCHEMA_CHANGE = 9 [(gogoproto.enumvalue_customname) = "TypeTypeSchemaChange"];
   STREAM_INGESTION = 10 [(gogoproto.enumvalue_customname) = "TypeStreamIngestion"];
   NEW_SCHEMA_CHANGE = 11 [(gogoproto.enumvalue_customname) = "TypeNewSchemaChange"];
+  MIGRATION = 12 [(gogoproto.enumvalue_customname) = "TypeMigration"];
 }
 
 message Job {

pkg/jobs/jobspb/wrap.go

@@ -30,6 +30,7 @@ var _ Details = CreateStatsDetails{}
 var _ Details = SchemaChangeGCDetails{}
 var _ Details = StreamIngestionDetails{}
 var _ Details = NewSchemaChangeDetails{}
+var _ Details = MigrationDetails{}
 
 // ProgressDetails is a marker interface for job progress details proto structs.
 type ProgressDetails interface{}
@@ -42,6 +43,7 @@ var _ ProgressDetails = CreateStatsProgress{}
 var _ ProgressDetails = SchemaChangeGCProgress{}
 var _ ProgressDetails = StreamIngestionProgress{}
 var _ ProgressDetails = NewSchemaChangeProgress{}
+var _ ProgressDetails = MigrationProgress{}
 
 // Type returns the payload's job type.
 func (p *Payload) Type() Type {
@@ -75,6 +77,8 @@ func DetailsType(d isPayload_Details) Type {
 		return TypeStreamIngestion
 	case *Payload_NewSchemaChange:
 		return TypeNewSchemaChange
+	case *Payload_Migration:
+		return TypeMigration
 	default:
 		panic(errors.AssertionFailedf("Payload.Type called on a payload with an unknown details type: %T", d))
 	}
@@ -109,6 +113,8 @@ func WrapProgressDetails(details ProgressDetails) interface {
 		return &Progress_StreamIngest{StreamIngest: &d}
 	case NewSchemaChangeProgress:
 		return &Progress_NewSchemaChange{NewSchemaChange: &d}
+	case MigrationProgress:
+		return &Progress_Migration{Migration: &d}
 	default:
 		panic(errors.AssertionFailedf("WrapProgressDetails: unknown details type %T", d))
 	}
@@ -138,6 +144,8 @@ func (p *Payload) UnwrapDetails() Details {
 		return *d.StreamIngestion
 	case *Payload_NewSchemaChange:
 		return *d.NewSchemaChange
+	case *Payload_Migration:
+		return *d.Migration
 	default:
 		return nil
 	}
@@ -167,6 +175,8 @@ func (p *Progress) UnwrapDetails() ProgressDetails {
 		return *d.StreamIngest
 	case *Progress_NewSchemaChange:
 		return *d.NewSchemaChange
+	case *Progress_Migration:
+		return *d.Migration
 	default:
 		return nil
 	}
@@ -209,6 +219,8 @@ func WrapPayloadDetails(details Details) interface {
 		return &Payload_StreamIngestion{StreamIngestion: &d}
 	case NewSchemaChangeDetails:
 		return &Payload_NewSchemaChange{NewSchemaChange: &d}
+	case MigrationDetails:
+		return &Payload_Migration{Migration: &d}
 	default:
 		panic(errors.AssertionFailedf("jobs.WrapPayloadDetails: unknown details type %T", d))
 	}
@@ -244,7 +256,7 @@ const (
 func (Type) SafeValue() {}
 
 // NumJobTypes is the number of jobs types.
-const NumJobTypes = 12
+const NumJobTypes = 13
 
 func init() {
 	if len(Type_name) != NumJobTypes {

pkg/migration/BUILD.bazel

@@ -12,6 +12,7 @@ go_library(
         "//pkg/clusterversion",
         "//pkg/kv",
         "//pkg/roachpb",
+        "//pkg/security",
         "//pkg/server/serverpb",
         "//pkg/util/log",
         "@com_github_cockroachdb_logtags//:logtags",

pkg/migration/migration.go

@@ -28,13 +28,14 @@ import (
 	"github.com/cockroachdb/cockroach/pkg/clusterversion"
 	"github.com/cockroachdb/cockroach/pkg/kv"
 	"github.com/cockroachdb/cockroach/pkg/roachpb"
+	"github.com/cockroachdb/cockroach/pkg/security"
 	"github.com/cockroachdb/cockroach/pkg/server/serverpb"
 	"github.com/cockroachdb/logtags"
 )
 
 // Manager coordinates long-running migrations.
 type Manager interface {
-	Migrate(ctx context.Context, from, to clusterversion.ClusterVersion) error
+	Migrate(ctx context.Context, user security.SQLUsername, from, to clusterversion.ClusterVersion) error
 }
 
 // Cluster abstracts a physical KV cluster and can be utilized by a long-runnng

pkg/migration/migrationjob/BUILD.bazel

@@ -0,0 +1,16 @@
+load("@io_bazel_rules_go//go:def.bzl", "go_library")
+
+go_library(
+    name = "migrationjob",
+    srcs = ["migration_job.go"],
+    importpath = "github.com/cockroachdb/cockroach/pkg/migration/migrationjob",
+    visibility = ["//visibility:public"],
+    deps = [
+        "//pkg/jobs",
+        "//pkg/jobs/jobspb",
+        "//pkg/migration",
+        "//pkg/migration/migrations",
+        "//pkg/settings/cluster",
+        "//pkg/sql",
+    ],
+)

pkg/migration/migrationjob/migration_job.go

@@ -0,0 +1,53 @@
+// Copyright 2018 The Cockroach Authors.
+//
+// Use of this software is governed by the Business Source License
+// included in the file licenses/BSL.txt.
+//
+// As of the Change Date specified in that file, in accordance with
+// the Business Source License, use of this software will be governed
+// by the Apache License, Version 2.0, included in the file
+// licenses/APL.txt.
+
+// Package migrationjob contains the jobs.Resumer implementation
+// used for long-running migrations.
+package migrationjob
+
+import (
+	"context"
+
+	"github.com/cockroachdb/cockroach/pkg/jobs"
+	"github.com/cockroachdb/cockroach/pkg/jobs/jobspb"
+	"github.com/cockroachdb/cockroach/pkg/migration"
+	"github.com/cockroachdb/cockroach/pkg/migration/migrations"
+	"github.com/cockroachdb/cockroach/pkg/settings/cluster"
+	"github.com/cockroachdb/cockroach/pkg/sql"
+)
+
+func init() {
+	jobs.RegisterConstructor(jobspb.TypeMigration, func(job *jobs.Job, settings *cluster.Settings) jobs.Resumer {
+		return &resumer{j: job}
+	})
+}
+
+type resumer struct {
+	j *jobs.Job
+}
+
+var _ jobs.Resumer = (*resumer)(nil)
+
+func (r resumer) Resume(ctx context.Context, execCtxI interface{}) error {
+	// TODO(ajwerner): add some check to see if we're done.
+	execCtx := execCtxI.(sql.JobExecContext)
+	pl := r.j.Payload()
+	cv := *pl.GetMigration().ClusterVersion
+	m, ok := migrations.GetMigration(cv)
+	if !ok {
+		return nil
+	}
+	return m.(*migration.KVMigration).Run(ctx, cv, execCtx.MigrationCluster())
+}
+
+// The long-running migration resumer has no reverting logic.
+func (r resumer) OnFailOrCancel(ctx context.Context, execCtx interface{}) error {
+	return nil
+}

pkg/migration/migrationmanager/BUILD.bazel

@@ -7,13 +7,21 @@ go_library(
     visibility = ["//visibility:public"],
     deps = [
         "//pkg/clusterversion",
+        "//pkg/jobs",
+        "//pkg/jobs/jobspb",
         "//pkg/kv",
         "//pkg/migration",
-        "//pkg/migration/migrationcluster",
+        "//pkg/migration/migrationjob",
         "//pkg/migration/migrations",
+        "//pkg/security",
         "//pkg/server/serverpb",
+        "//pkg/sql/protoreflect",
+        "//pkg/sql/sem/tree",
+        "//pkg/sql/sqlutil",
         "//pkg/util/log",
+        "@com_github_cockroachdb_errors//:errors",
         "@com_github_cockroachdb_logtags//:logtags",
+        "@com_github_cockroachdb_redact//:redact",
     ],
 )
 
@@ -26,6 +34,7 @@ go_test(
     deps = [
         "//pkg/base",
         "//pkg/clusterversion",
+        "//pkg/jobs",
         "//pkg/kv/kvserver/batcheval",
         "//pkg/kv/kvserver/liveness",
         "//pkg/migration",
@@ -35,11 +44,13 @@ go_test(
         "//pkg/security/securitytest",
         "//pkg/server",
         "//pkg/settings/cluster",
+        "//pkg/sql/sqlutil",
         "//pkg/testutils",
         "//pkg/testutils/serverutils",
         "//pkg/testutils/testcluster",
         "//pkg/util/leaktest",
         "//pkg/util/log",
+        "//pkg/util/tracing",
         "@com_github_cockroachdb_errors//:errors",
         "@com_github_stretchr_testify//require",
     ],