Oracle归档日志所在目录时间不对&&Oracle集群日志时间显示错误

Oracle归档日志所在目录时间不对&&Oracle集群日志时间显示错误

前言

这个问题在18年的时候遇到了，基本不注意并且集群或者数据库运行正常是很难注意到的。

忘记当时怎么发现的了，最近拿出来重现一下问题。

具体描述一下问题：

比如当前时间是2020年07月03日0点，数据库对redo log发起归档。

在OMF下，归档会按照日期进行分类归档。

但是发起归档之后，新的归档文件会放在的20200702的目录中。

另外，db和asm和集群的警告日志的时间也和操作系统的date时间不一致。

现象

在我的测试环境中，关闭所有节点的集群之后，更改时间：

[root@rac1 ~]# date -s "2020-07-03 00:00:00"
Fri Jul   :: CST 

然后，启动所有节点的集群，可以依次观察

集群的警告日志，grid用户下，$ORACLE_HOME/log/rac1/alertrac1.log

-- ::26.667:
[ohasd()]CRS-:The OLR service started on node rac1.
-- ::26.675:
[ohasd()]CRS-:Oracle High Availability Service started on node rac1.
-- ::26.676:
[ohasd()]CRS-:location: /etc/oracle/lastgasp has  reboot advisory log files,  were announced and  errors occurred
-- ::30.080:
[/u01/app/11.2./grid/bin/orarootagent.bin()]CRS-:Cannot get GPnP profile. Error CLSGPNP_NO_DAEMON (GPNPD daemon is not running).
-- ::34.563:
[gpnpd()]CRS-:GPNPD started on node rac1.
-- ::36.992:
[cssd()]CRS-:CSSD daemon is started in clustered mode
-- ::38.798:
[ohasd()]CRS-:Resource state recovery not attempted for 'ora.diskmon' as its target state is OFFLINE
-- ::01.319:
[cssd()]CRS-:Lease acquisition for node rac1 number  completed
-- ::02.612:
[cssd()]CRS-:CSSD voting file is online: /dev/mapper/asm-ocr02; details in /u01/app/11.2./grid/log/rac1/cssd/ocssd.log.
-- ::02.620:
[cssd()]CRS-:CSSD voting file is online: /dev/mapper/asm-ocr03; details in /u01/app/11.2./grid/log/rac1/cssd/ocssd.log.
-- ::02.623:
[cssd()]CRS-:CSSD voting file is online: /dev/mapper/asm-ocr01; details in /u01/app/11.2./grid/log/rac1/cssd/ocssd.log.
-- ::15.917:
[cssd()]CRS-:CSSD Reconfiguration complete. Active nodes are rac1 rac2 .
-- ::18.061:
[ctssd()]CRS-:The Cluster Time Synchronization Service on host rac1 is in observer mode.
-- ::18.421:
[ctssd()]CRS-:The new Cluster Time Synchronization Service reference node is host rac1.
-- ::18.422:
[ctssd()]CRS-:The Cluster Time Synchronization Service started on host rac1.
-- ::20.023:
[ohasd()]CRS-:Resource state recovery not attempted for 'ora.diskmon' as its target state is OFFLINE
[client()]CRS-:-Jul- : ACFS-: Checking for existing ADVM/ACFS installation.
[client()]CRS-:-Jul- : ACFS-: Validating ADVM/ACFS installation files for operating system.
[client()]CRS-:-Jul- : ACFS-: Verifying ASM Administrator setup.
[client()]CRS-:-Jul- : ACFS-: Loading installed ADVM/ACFS drivers.
[client()]CRS-:-Jul- : ACFS-: Loading 'oracleoks.ko' driver.
[client()]CRS-:-Jul- : ACFS-: Loading 'oracleadvm.ko' driver.
[client()]CRS-:-Jul- : ACFS-: Loading 'oracleacfs.ko' driver.
........

ASM实例警告日志，

Thu Jul  ::
NOTE: No asm libraries found in the system
NOTE: No asm libraries found in the system
MEMORY_TARGET defaulting to .
* instance_number obtained from CSS = , checking for the existence of node ...
* node  does not exist. instance_number =
Starting ORACLE instance (normal)
WARNING: You are trying to use the MEMORY_TARGET feature. This feature requires the /dev/shm file system to be mounted for at least  bytes. /dev/shm is either not mounted or is mounted with available space less than this size. Please fix this so that MEMORY_TARGET can work as expected. Current available is  and used is  bytes. Ensure that the mount point is /dev/shm for this directory.
LICENSE_MAX_SESSION =
LICENSE_SESSIONS_WARNING =
Initial number of CPU is
Number of processor cores in the system is
Number of processor sockets in the system is
Private Interface 'eth1:1' configured from GPnP for use as a private interconnect.
  [name='eth1:1', type=, ip=169.254.146.181, mac=-0c----8b, net=169.254.0.0/, mask=255.255.0.0, use=haip:cluster_interconnect/]
Public Interface 'eth0' configured from GPnP for use as a public interface.
  [name='eth0', type=, ip=192.168.180.100, mac=-0c----, net=192.168.180.0/, mask=255.255.255.0, use=public/]
Picked latch-free SCN scheme
Using LOG_ARCHIVE_DEST_1 parameter default value as /u01/app/11.2./grid/dbs/arch
Autotune of undo retention is turned on.
LICENSE_MAX_USERS =
SYS auditing is disabled
NOTE: Volume support  enabled
Starting up:
Oracle Database 11g Enterprise Edition Release 11.2.0.4. - 64bit Production
With the Real Application Clusters and Automatic Storage Management options.
ORACLE_HOME = /u01/app/11.2./grid
System name:    Linux
Node name:      rac1.example.com
Release:        2.6.-.el6.x86_64
Version:        # SMP Wed Jul  :: EDT
Machine:        x86_64
VM name:        VMWare Version:
Using parameter settings in server-side spfile +OCR/rac-cluster/asmparameterfile/registry.253.1034884609

DB的警告日志，

Thu Jul  ::
Adjusting the default value of parameter parallel_max_servers
from  to  due to the value of parameter processes ()
Starting ORACLE instance (normal)
LICENSE_MAX_SESSION =
LICENSE_SESSIONS_WARNING =
Initial number of CPU is
Number of processor cores in the system is
Number of processor sockets in the system is
Private Interface 'eth1:1' configured from GPnP for use as a private interconnect.
  [name='eth1:1', type=, ip=169.254.146.181, mac=-0c----8b, net=169.254.0.0/, mask=255.255.0.0, use=haip:cluster_interconnect/]
Public Interface 'eth0' configured from GPnP for use as a public interface.
  [name='eth0', type=, ip=192.168.180.100, mac=-0c----, net=192.168.180.0/, mask=255.255.255.0, use=public/]
Public Interface 'eth0:2' configured from GPnP for use as a public interface.
  [name='eth0:2', type=, ip=192.168.180.111, mac=-0c----, net=192.168.180.0/, mask=255.255.255.0, use=public/]
Picked latch-free SCN scheme
Autotune of undo retention is turned on.
LICENSE_MAX_USERS =
SYS auditing is disabled
Starting up:
Oracle Database 11g Enterprise Edition Release 11.2.0.4. - 64bit Production
With the Partitioning, Real Application Clusters, OLAP, Data Mining
and Real Application Testing options.
ORACLE_HOME = /u01/app/oracle/product/11.2./db_1
System name:    Linux
Node name:      rac1.example.com
Release:        2.6.-.el6.x86_64

可以看到，日志开头全都是2020年07月02日的9点初。

我们再看看归档的情况，

在此之前，先删除之前的所有归档便于观察。

[oracle@rac2 ~]$ rman target /

Recovery Manager: Release 11.2.0.4.0 - Production on Fri Jul 3 00:20:01 2020

Copyright (c) 1982, 2011, Oracle and/or its affiliates.  All rights reserved.

connected to target database: ZKM (DBID=4210149549)

RMAN> delete force noprompt archivelog all;

using target database control file instead of recovery catalog
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=200 instance=zkm2 device type=DISK
List of Archived Log Copies for database with db_unique_name ZKM
=====================================================================

Key     Thrd Seq     S Low Time
------- ---- ------- - ---------
91      1    56      A 02-JUL-20
        Name: +ARCH/zkm/archivelog/2020_07_03/thread_1_seq_56.303.1044695873

92      2    21      A 02-JUL-20
        Name: +ARCH/zkm/archivelog/2020_07_02/thread_2_seq_21.299.1044695875

deleted archived log
archived log file name=+ARCH/zkm/archivelog/2020_07_03/thread_1_seq_56.303.1044695873 RECID=91 STAMP=1044749873
deleted archived log
archived log file name=+ARCH/zkm/archivelog/2020_07_02/thread_2_seq_21.299.1044695875 RECID=92 STAMP=1044695874
Deleted 2 objects

其中一个节点发出命令归档redo log，

SYS@zkm1> alter session set nls_date_format='yyyy-mm-dd hh24:mi:ss';

Session altered.

SYS@zkm1>
SYS@zkm1> select sysdate from dual;

SYSDATE
------------------------
2020-07-03 00:20:27

SYS@zkm1> alter system archive log current;

System altered.

观察警告日志，显示的时间是“Thu Jul 02 09:20:32 2020”。

Thu Jul 02 09:20:32 2020
Thread 1 advanced to log sequence 58 (LGWR switch)
  Current log# 2 seq# 58 mem# 0: +DATA/zkm/onlinelog/group_2.258.1034889845
  Current log# 2 seq# 58 mem# 1: +ARCH/zkm/onlinelog/group_2.258.1034889845
Thu Jul 02 09:20:32 2020
Archived Log entry 93 added for thread 1 sequence 57 ID 0xfaf1eead dest 1:

进去ASM磁盘查看文件情况，

ASMCMD> ls -l
Type  Redund  Striped  Time             Sys  Name
                                        Y    2020_07_02/
ASMCMD> ls -l *
Type        Redund  Striped  Time             Sys  Name
ARCHIVELOG  UNPROT  COARSE   JUL 02 09:00:00  Y    thread_1_seq_57.299.1044696033
ARCHIVELOG  UNPROT  COARSE   JUL 02 09:00:00  Y    thread_2_seq_22.303.1044696033
ASMCMD> pwd
+arch/zkm/ARCHIVELOG

至此，问题完全重现成功。

原因

不知道各位发现没有，实际上最开始启动集群（crsctl start crs）那一瞬间（大概是2020-07-03 00:00:00多，刚改完时间不到1分钟），

集群的警告日志最开始出现的时间为“2020-07-02 09:00:26.667:”，基本相差15小时。

因此很容易也可以猜出，是由于时区的原因。

系统的时区为：

[grid@rac1 ~]$ cat /etc/sysconfig/clock
# The time zone of the system is defined by the contents of /etc/localtime.
# This file is only for evaluation by system-config-date, do not rely on its
# contents elsewhere.
ZONE="Asia/Shanghai"

那么，集群应该也有自己的配置文件对应的时区，也就是集群不依赖OS的时区（-_-||）。

mos可以找到这么一篇文章：

How To Change Timezone for Grid Infrastructure (Doc ID 1209444.1)