OracleExadataDatabaseMachineGridInfrastructureUpgradeto11.2.0.2failswithASMORA-4031
适用于:
OracleExadata硬件-版:11.2.0.2——释放到11.2.0.2:11.2到11.2之间波动
信息在这个文件适用于其他平台。
症状
在网格基础设施升级到11.2.0.2 11.2.0.1、运行rootupgrade。上海第一节点可能会失败。ASM警报日志显示:
ORA-04031:unabletoallocate760bytesofsharedmemory("sharedpool","unknownobject","KKSSP^3","kglss")
rootupgrade.sh承宪退出声明该脚本ASM资源不能开始:
开始开发资源”ora.Asm“失败
其他节点11.2.0.1仍然运行部件。
这是的仍在等待确认rootupgrade.sh已经运行在所有节点。
引起
默认的部署甲骨文数据库Exadata许多机器包括设置学会情况下这样的:sga_max_size = 1 g。这个设置是不足以进行升级。这被记录为错误10386429。未出版
如果你按照这些步骤在文档1315926.1之后(特别是设置和复位memory_target sga_max_size)在开始升级或至少在rootupgrade运行。上海,这个问题描述在这本笔记不应该发生,这些回收脚步必不被需要。
例如:
altersystemsetmemory_target=1Gscope=spfile;
altersystemresetsga_targetscope=spfile;
altersystemresetsga_max_sizescope=spfile;
altersystemresetpga_aggregate_targetscope=spfile;
altersystemresetmemory_max_targetscope=spfile;
溶液
假设
圣谕师普遍安装(是的)仍在运转,等待rootupgrade证实。上海已经运行在所有节点。让它一直运转,不要点击OK呢。
这些要求是不跟随在开始升级(特别是设置和复位memory_target sga_max_size)-看”导致“节以上。
rootupgrade的。承宪脚本失败在第一个节点的地方跑去。其它节点仍然运行11.2.0.1释放。本程序仅作为测试数据库Oracle Exadata机,没有其他来源的版本进行了试验。在第一个节点故障信息,包括这个样子的输出:
Startofresource"ora.asm"failed
CRS-2672:Attemptingtostart'ora.drivers.acfs'on'burexdk50db01'
CRS-2676:Startof'ora.drivers.acfs'on'dm01db01'succeeded
CRS-2672:Attemptingtostart'ora.asm'on'dm01db01'
CRS-5017:Theresourceaction"ora.asmstart"encounteredthefollowingerror:
ORA-03113:end-of-fileoncommunicationchannel
ProcessID:0
SessionID:0Serialnumber:0CRS-2674:Startof'ora.asm'on'dm01db01'failed
CRS-2679:Attemptingtoclean'ora.asm'on'dm01db01'
CRS-2681:Cleanof'ora.asm'on'dm01db01'succeeded
CRS-2673:Attemptingtostop'ora.drivers.acfs'on'dm01db01'
CRS-2677:Stopof'ora.drivers.acfs'on'dm01db01'succeeded
CRS-4000:CommandStartfailed,orcompletedwitherrors.
FailedtostartOracleClusterwarestack
FailedtostartASMat/u01/app/11.2.0.2/grid/crs/install/crsconfig_lib.pmline1051.
/u01/app/11.2.0.2/grid/perl/bin/perl-I/u01/app/11.2.0.2/grid/perl/lib-I/u01/app/11.2.0.2/grid/crs/install/u01/app/11.2.0.2/grid/crs/install/rootcrs.plexecutionfailed
你没有修改了ASM内存配置参数的默认的部署,或者如果你有,sga_max_size = 1 g的遗迹。
如果你点击可以在rootupgrade.sh完之前就可以.第一,那么你可能不会严格的遵循这张纸条。相反,遵循“失败在969254.1以下文件rootupgrade。上海“开始第一步。
恢复步骤完成升级
第一次(在rootupgrade节点。承宪失败发生),停止使用11.2.0.2过程11.2.0.2和胃肠道双星如图所示:
[root@dscgigdb05grid]#/u01/app/11.2.0.2/grid/bin/crsctlstopcrs
CRS-2791:StartingshutdownofOracleHighAvailabilityServices-managedresourceson'dscgigdb05'
CRS-2673:Attemptingtostop'ora.crf'on'dscgigdb05'
CRS-2673:Attemptingtostop'ora.ctssd'on'dscgigdb05'
CRS-2673:Attemptingtostop'ora.cluster_interconnect.haip'on'dscgigdb05'
CRS-2673:Attemptingtostop'ora.mdnsd'on'dscgigdb05'
CRS-2677:Stopof'ora.crf'on'dscgigdb05'succeeded
CRS-2677:Stopof'ora.mdnsd'on'dscgigdb05'succeeded
CRS-2677:Stopof'ora.cluster_interconnect.haip'on'dscgigdb05'succeeded
CRS-2677:Stopof'ora.ctssd'on'dscgigdb05'succeeded
CRS-2673:Attemptingtostop'ora.cssd'on'dscgigdb05'
CRS-2677:Stopof'ora.cssd'on'dscgigdb05'succeeded
CRS-2673:Attemptingtostop'ora.gipcd'on'dscgigdb05'
CRS-2673:Attemptingtostop'ora.diskmon'on'dscgigdb05'
CRS-2677:Stopof'ora.diskmon'on'dscgigdb05'succeeded
CRS-2677:Stopof'ora.gipcd'on'dscgigdb05'succeeded
CRS-2673:Attemptingtostop'ora.gpnpd'on'dscgigdb05'
CRS-2677:Stopof'ora.gpnpd'on'dscgigdb05'succeeded
CRS-2793:ShutdownofOracleHighAvailabilityServices-managedresourceson'dscgigdb05'hascompleted
CRS-4133:OracleHighAvailabilityServiceshasbeenstopped.
[root@dscgigdb05grid]#
在一个幸存的节点(仍在运行的11.2.0.1,不升级,还连接到学会。注意/及其他/oratab文件就已经被更新为ASM实例,所以如果你使用oraenv设定你的环境,你可能需要手动修改/及其他/oratab或手动设置一个环境变量为了要连接。接过(/sysasm),运行:
$sqlplus/nolog###environmentissetforASMinstance
SQL>connect/assysasm
Connected.
SQL>altersystemstoprollingmigration;
System altered.
SQL> select sys_context('sys_cluster_properties','cluster_state') from dual;
SYS_CONTEXT('SYS_CLUSTER_PROPERTIES','CLUSTER_STATE')
--------------------------------------------------------------------------------
正常
如果cluster_state不正常,那么联系Oracle的支持和停止这些步骤。否则,继续下列步骤在同一sqlplus会议。
SQL> alter system set memory_target=1g scope=spfile;
System altered.
SQL> show parameter sga
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
lock_sga boolean FALSE
sga_max_size big integer 1G
sga_target big integer 0
SQL> alter system reset sga_max_size scope=spfile sid='*';
System altered.
注:以下有些改变系统可能会产生错误陈述参数不设置。这些错误是可以忽略的。
SQL> alter system reset sga_target scope=spfile;
System altered.
SQL> alter system reset sga_max_size scope=spfile;
System altered.
SQL> alter system reset pga_aggregate_target scope=spfile;
System altered.
SQL> alter system reset memory_max_target scope=spfile;
System altered.
SQL> ALTER SYSTEM START ROLLING MIGRATION TO '11.2.0.2.0';
System altered.
这些步骤只需要运行从一个保存至今的ASM情况下,不是每个实例。无零件需要重新启动于这一点。
第一次在rootupgrade节点。承宪没有以前,rootupgrade重演。承宪像以前一样:
[root@dscgigdb05~]#/u01/app/11.2.0.2/grid/rootupgrade.sh
(output deleted)
The rootupgrade.sh output may include the following message near the start of the output (just after the oraenv, coraenv prompts):
Failed to add (property/value):('OLD_OCR_ID/'-1') for checkpoint:ROOTCRS_OLDHOMEINFO.Error code is 256
This is Bug:10056593, the workaround is to ignore the message - it will be fixed in 11.2.0.3 and won't harm the upgrade.
The last line of output should look like this:
Configure Oracle Grid Infrastructure for a Cluster ... succeeded
在这一点上,第一个节点是成功的。然而,如果你试图抑制状态从加强节点,你可能会遇到未公开的错误9615596:
[root@dscgigdb05 ~]# /u01/app/11.2.0.2/grid/bin/crsctl stat res -t
CRS-601: Internal error
RC: 5, File: clsEntityDisplay.cpp, Line: 265
CRS-4000: Command Status failed, or completed with errors.
[root@dscgigdb05 ~]#
这是未公开的错误9615596。错误就会消失在所有节点完成他们的升级。忽略错误并继续。
然而,从其他节点(仍在11.2.0.1),运行相同的命令GRID_HOME 11.2.0.1看起来很普通。
其次,rootupgrade运行。承宪在其他节点和确认他们的成功。你应该看看这最后的输出。
Configure Oracle Grid Infrastructure for a Cluster ... succeeded
利用网格基础设施从11.2.0.2二进制文件,所有节点现在应该报告从正常状态的命令:crsctlstatres-t
早在安装程序(rootupgrade.shy运行和等待。承宪脚本运行),点击行根脚本安装提示符窗口。邮政升级步骤应该安装成功完成。Errors
ORA-4031;ORA-3113