关于并行重做的消息

Question

Parallel redo is shutdown for database '' with worker pool size [2].

Parallel redo is started for database '' with worker pool size [2].

Starting up database ''

我在客户端 PC 上的 Windows 事件查看器（事件 ID 49930 或 17137）中经常看到这种情况。这是一台普通的 Windows PC，而不是服务器，带有 SQL Server 2017 Express Edition 和 SSMS 的默认实例。

什么是并行重做？

这是错误日志的输出：

2019-05-28 12:23:03.360 spid16s A self-generated certificate was successfully loaded for encryption.
2019-05-28 12:23:03.360 spid16s Server is listening on [ 'any' <ipv6> 1433].
2019-05-28 12:23:03.360 spid16s Server is listening on [ 'any' <ipv4> 1433].
2019-05-28 12:23:03.360 spid16s Server is listening on [ 'any' <ipv6> 50374].
2019-05-28 12:23:03.360 spid16s Server is listening on [ 'any' <ipv4> 50374].
2019-05-28 12:23:03.360 spid16s Server local connection provider is ready to accept connection on [ \\.\pipe\SQLLocal\FOLLOWIN ].
2019-05-28 12:23:03.360 spid16s Server named pipe provider is ready to accept connection on [ \\.\pipe\MSSQL$FOLLOWIN\sql\query ].
2019-05-28 12:23:03.360 spid16s Dedicated administrator connection support was not started because it is disabled on this edition of SQL Server. If you want to use a dedicated administrator connection, restart SQL Server using the trace flag 7806. This is an informational message only. No user action is required.
2019-05-28 12:23:03.360 Server  SQL Server is attempting to register a Service Principal Name (SPN) for the SQL Server service. Kerberos authentication will not be possible until a SPN is registered for the SQL Server service. This is an informational message. No user action is required.
2019-05-28 12:23:03.360 spid16s SQL Server is now ready for client connections. This is an informational message; no user action is required.
2019-05-28 12:23:03.370 Server  The SQL Server Network Interface library could not register the Service Principal Name (SPN) [ MSSQLSvc/pcFoo:FOLLOWIN ] for the SQL Server service. Windows return code: 0xffffffff, state: 63. Failure to register a SPN might cause integrated authentication to use NTLM instead of Kerberos. This is an informational message. Further action is only required if Kerberos authentication is required by authentication policies and if the SPN has not been manually registered.
2019-05-28 12:23:03.370 Server  The SQL Server Network Interface library could not register the Service Principal Name (SPN) [ MSSQLSvc/pcFoo:1433 ] for the SQL Server service. Windows return code: 0xffffffff, state: 63. Failure to register a SPN might cause integrated authentication to use NTLM instead of Kerberos. This is an informational message. Further action is only required if Kerberos authentication is required by authentication policies and if the SPN has not been manually registered.
2019-05-28 12:23:03.430 spid17s A new instance of the full-text filter daemon host process has been successfully started.
2019-05-28 12:23:03.520 spid12s Starting up database 'mssqlsystemresource'.
2019-05-28 12:23:03.520 spid21s Starting up database 'Foobar'.
2019-05-28 12:23:03.520 spid20s Starting up database 'msdb'.
2019-05-28 12:23:03.530 spid12s The resource database build version is 14.00.3162. This is an informational message only. No user action is required.
2019-05-28 12:23:03.550 spid12s Synchronize Database 'master' (1) with Resource Database.
2019-05-28 12:23:03.550 spid21s Parallel redo is started for database 'Foobar' with worker pool size [2].
2019-05-28 12:23:03.560 spid12s Starting up database 'model'.
2019-05-28 12:23:03.560 spid21s Parallel redo is shutdown for database 'Foobar' with worker pool size [2].
2019-05-28 12:23:03.580 spid12s Synchronize Database 'model' (3) with Resource Database.
2019-05-28 12:23:03.590 spid12s Polybase feature disabled.
2019-05-28 12:23:03.590 spid12s Clearing tempdb database.
2019-05-28 12:23:03.670 spid12s Starting up database 'tempdb'.
2019-05-28 12:23:03.710 spid20s Synchronize Database 'msdb' (4) with Resource Database.
2019-05-28 12:23:03.710 spid21s Synchronize Database 'Foobar' (5) with Resource Database.
2019-05-28 12:23:03.710 spid22s The Service Broker endpoint is in disabled or stopped state.
2019-05-28 12:23:03.710 spid22s The Database Mirroring endpoint is in disabled or stopped state.
2019-05-28 12:23:03.730 spid22s Service Broker manager has started.
2019-05-28 12:23:03.730 spid6s  Recovery is complete. This is an informational message only. No user action is required.
2019-05-28 12:28:21.180 spid51  Attempting to load library 'xplog70.dll' into memory. This is an informational message only. No user action is required.
2019-05-28 12:28:21.180 spid51  Using 'xplog70.dll' version '2017.140.3162' to execute extended stored procedure 'xp_msver'. This is an informational message only; no user action is required.
2019-05-28 12:28:21.320 spid51  Starting up database 'Foobar'.
2019-05-28 12:28:21.340 spid51  Parallel redo is started for database 'Foobar' with worker pool size [2].
2019-05-28 12:28:21.350 spid51  Parallel redo is shutdown for database 'Foobar' with worker pool size [2].
2019-05-28 12:28:21.360 spid51  Starting up database 'Foo'.
2019-05-28 12:28:21.380 spid51  Parallel redo is started for database 'Foo' with worker pool size [2].
2019-05-28 12:28:21.400 spid51  Parallel redo is shutdown for database 'Foo' with worker pool size [2].
2019-05-28 12:28:21.410 spid51  Synchronize Database 'Foo' (6) with Resource Database.
2019-05-28 14:27:21.430 spid59  Starting up database 'Foo'.
2019-05-28 14:27:21.450 spid59  Parallel redo is started for database 'Foo' with worker pool size [2].
2019-05-28 14:27:21.470 spid59  Parallel redo is shutdown for database 'Foo' with worker pool size [2].
2019-05-28 14:27:21.480 spid59  Starting up database 'Foobar'.
2019-05-28 14:27:21.500 spid59  Parallel redo is started for database 'Foobar' with worker pool size [2].
2019-05-28 14:27:21.510 spid59  Parallel redo is shutdown for database 'Foobar' with worker pool size [2].
2019-05-28 14:27:38.330 spid59  Starting up database 'Foobar'.
2019-05-28 14:27:38.360 spid59  Parallel redo is started for database 'Foobar' with worker pool size [2].
2019-05-28 14:27:38.370 spid59  Parallel redo is shutdown for database 'Foobar' with worker pool size [2].
2019-05-28 14:27:40.010 spid59  Starting up database 'Foo'.
2019-05-28 14:27:40.050 spid59  Parallel redo is started for database 'Foo' with worker pool size [2].
2019-05-28 14:27:40.080 spid59  Parallel redo is shutdown for database 'Foo' with worker pool size [2].
2019-05-28 14:28:04.160 spid59  Starting up database 'Foo'.
2019-05-28 14:28:04.180 spid59  Parallel redo is started for database 'Foo' with worker pool size [2].
2019-05-28 14:28:04.210 spid59  Parallel redo is shutdown for database 'Foo' with worker pool size [2].
2019-05-28 14:28:10.620 spid52  Attempting to load library 'xpstar.dll' into memory. This is an informational message only. No user action is required.
2019-05-28 14:28:10.690 spid52  Using 'xpstar.dll' version '2017.140.3162' to execute extended stored procedure 'xp_instance_regread'. This is an informational message only; no user action is required.
2019-05-28 14:28:14.220 spid52  Starting up database 'Foo'.
2019-05-28 14:28:14.240 spid52  Parallel redo is started for database 'Foo' with worker pool size [2].
2019-05-28 14:28:14.270 spid52  Parallel redo is shutdown for database 'Foo' with worker pool size [2].
2019-05-28 14:28:14.410 Backup  Database backed up. Database: Foo, creation date(time): 2019/05/08(15:06:00), pages dumped: 2555, first LSN: 48:9778:37, last LSN: 48:9795:1, number of dump devices: 1, device information: (FILE=2, TYPE=DISK: {'C:\Users\Joe\Documents\Backup Followin\Foo_FULL_05272019_235209.BAK'}). This is an informational message only. No user action is required.
2019-05-28 14:28:14.420 Backup  BACKUP DATABASE successfully processed 2546 pages in 0.112 seconds (177.529 MB/sec).
2019-05-28 14:28:19.120 spid53  Starting up database 'Foo'.
2019-05-28 14:28:19.140 spid53  Parallel redo is started for database 'Foo' with worker pool size [2].
2019-05-28 14:28:19.160 spid53  Parallel redo is shutdown for database 'Foo' with worker pool size [2].
2019-05-28 14:28:20.610 spid59  Starting up database 'Foo'.
2019-05-28 14:28:20.650 spid59  Parallel redo is started for database 'Foo' with worker pool size [2].
2019-05-28 14:28:20.670 spid59  Parallel redo is shutdown for database 'Foo' with worker pool size [2].
2019-05-28 14:28:31.030 spid54  Using 'dbghelp.dll' version '4.0.5'

0.3162.1 (X64)

May 15 2019 19:14:30

Copyright (C) 2017 Microsoft Corporation

Express Edition (64-bit) on Windows 10 Pro 10.0 <X64> (Build 17763: )

Answer 1

这是 SQL Server 2017 中添加的与可用性组的并行重做相关的信息性日志消息：

SQL Server 2016/2017：可用性组辅助副本重做模型和性能

您看到的行为（重做开始并立即停止）是正常的。这是因为数据库没有参与可用性组。当一个数据库启动时，该数据库的并行重做线程启动，然后意识到没有可用组，然后关闭。

您可以通过打开记录和支持的全局跟踪标志 3459 来禁用并行重做，然后使数据库脱机和联机来证明这一点。您将不会再看到这些消息。

请参阅我的错误日志的屏幕截图。

我启用了跟踪标志，关闭/联机数据库：没有消息

我禁用了跟踪标志，关闭/联机数据库：返回的消息

除非您打开该跟踪标志，否则您将在数据库联机时看到这些消息（因为服务器重新启动，或者数据库专门脱机然后重新联机，以及其他类似的情况）。但是，它们并不表示存在问题。

如果您的数据库正在参与 AG，并且您整天看到这些“重做已关闭”消息，则可能是因为AUTO_CLOSE已打开（Express Edition 上用户数据库的默认设置）。考虑将该设置更改为关闭，这通常被认为是“最佳实践”：

ALTER DATABASE YourDatabaseName SET AUTO_CLOSE OFF;

注：这些消息不涉及崩溃恢复的是在数据库启动时发生“重做”部分。