Oracle 19c RAC 遇到的几个问题

JiekeXu之路

发布于 2022-03-31 20:42:26

9720

发布于 2022-03-31 20:42:26

大家好，我是 JiekeXu,很高兴又和大家见面了,今天和大家一起来看看 Oracle 19c RAC 遇到的几个问题

Oracle19c 作为长期支持的大版本，是很多公司和个人选择的主流数据库版本，很多公司新上线的系统也都是以 19C 为主，也有很多企业渐渐地迁移数据库到19C，11204 版本已经逐渐退出了舞台。下面对 Oracle 19C版本遇到的几个最主要的问题做个简单介绍。

1、Fock() 炸弹

19 RAC 使用 srvctl 无法正常启动，reboot 重启也不会随着 OS 启动而启动数据库实例。

现象如下图：

Alert 日志报错

对应 trace

猜测可能是资源限制问题，但查了很多都是正常。

ipcs –ma 查看 Oracle 共享内存段

ipcs -q　

对应时间段的 OS message

Grid 用户资源限制情况，基本上都是无限制。

Oracle 用户资源限制情况，基本上都是无限制

使用 strace 跟踪启动过程

strace–o /tmp/strace2.log srvctl start database -d teststb –i teststb2

strace–o /tmp/strace1.log srvctl start database -d teststb -i teststb1

其他资源限制查看

cat /etc/security/limits.conf

cat /etc/systemd/system.conf

将此参数打开修改为 infinity 然后重启主机则恢复正常。

cat /etc/systemd/system.conf|grepDefaultTasksMax
#DefaultTasksMax=512
DefaultTasksMax=infinity

Database And ASM InstanceOra-27300 OS System Dependent Operation Fork Failed With Status 11 (Doc ID2331884.1)

systemd limited maximumnumber of tasks that may be created on the node.This setting will alsoaffect maxpid value on the OS.

主要是MAX_PID

修改DefaultTasksMax

从SLES 12 SP2开始引入了PID cgroup controller，限制fork()的并发数，避免fork()炸弹。由配置参数DefaultTasksMax控制，缺省值为512，该值比较小，不适用于数据库运行的环境，需要修改为大于等于65536，在此修改为infinity（无限制）。

如果在DBCA期间抛出如下图所示错误，则可能是由于DefaultTasksMax值过小导致的。为了避免DBCA出现错误，提前修改该参数值。

解决办法：

1）、查看当前值：

test1:~ #systemctl show --property DefaultTasksMax

DefaultTasksMax=512

2）、编辑/etc/systemd/system.conf，取消DefaultTasksMax该行注释，并修改值为"infinity".

# vi /etc/systemd/system.conf

DefaultTasksMax=infinity

3）、reboot OS

4）、查看修改后的值

oracle@test1:~>systemctl show --property DefaultTasksMax

DefaultTasksMax=18446744073709551615

2、客户端问题 ORA-28040

前置条件：已安装Oracle11g客户端，配置好环境变量，用PL/SQL Developer登录数据库

解决方案：

在oracle 19C服务器端oracle用户下：

cd $ORACLE_HOME/network/admin目录下 新建文件 sqlnet.ora
vi sqlnet.ora
SQLNET.ALLOWED_LOGON_VERSION_SERVER=10;
SQLNET.ALLOWED_LOGON_VERSION_CLIENT=10;

②在服务器端，管理员身份登录，重新修改密码。

sqlplus / as sysdba
alter session set container=服务名; --如果使用 pdb 需要切换到 pdb 修改。
alter user 用户名 identified by 密码;

注意：配置完服务器端的sqlnet.ora文件后，务必要重新修改密码，否则仍登录失败

3、打补丁报错

打补丁报错 Prerequisite check "CheckActiveFilesAndExecutables" failed。

testrac1:~# /app/app/19.0.0.0/grid/OPatch/opatchauto apply /app/soft/32226239

OPatchautosession is initiated at Thu Mar  415:27:50 2022

Systeminitialization log file is /app/app/19.0.0.0/grid/cfgtoollogs/opatchautodb/systemconfig2022-03-04_03-27-53PM.log.

Sessionlog file is /app/app/19.0.0.0/grid/cfgtoollogs/opatchauto/opatchauto2022-03-04_03-28-19PM.log
Theid for this session is U4JF
ExecutingOPatch prereq operations to verify patch applicability on home/app/app/oracle/product/19.0.0.0/dbhome_1
ExecutingOPatch prereq operations to verify patch applicability on home/app/app/19.0.0.0/grid
Patchapplicability verified successfully on home /app/app/19.0.0.0/grid
Patchapplicability verified successfully on home/app/app/oracle/product/19.0.0.0/dbhome_1
Executingpatch validation checks on home /app/app/19.0.0.0/grid
Patchvalidation checks successfully completed on home /app/app/19.0.0.0/grid
Executingpatch validation checks on home /app/app/oracle/product/19.0.0.0/dbhome_1
Patchvalidation checks successfully completed on home/app/app/oracle/product/19.0.0.0/dbhome_1

VerifyingSQL patch applicability on home /app/app/oracle/product/19.0.0.0/dbhome_1
"/bin/sh-c 'cd /app/app/oracle/product/19.0.0.0/dbhome_1;ORACLE_HOME=/app/app/oracle/product/19.0.0.0/dbhome_1 ORACLE_SID=piccyx1/app/app/oracle/product/19.0.0.0/dbhome_1/OPatch/datapatch -prereq-verbose'" command failed with errors. Please refer to logs for more details.SQL changes, if any, can be analyzed by manually retrying the same command.

SQLpatch applicability verified successfully on home/app/app/oracle/product/19.0.0.0/dbhome_1

Preparingto bring down database service on home /app/app/oracle/product/19.0.0.0/dbhome_1
Successfullyprepared home /app/app/oracle/product/19.0.0.0/dbhome_1 to bring down databaseservice

Bringingdown CRS service on home /app/app/19.0.0.0/grid
Prepatchoperation log file location: /app/app/grid/crsdata/testrac1/crsconfig/crs_prepatch_apply_inplace_testrac1_2022-03-04_03-29-01PM.log
CRSservice brought down successfully on home /app/app/19.0.0.0/grid

Performingprepatch operation on home /app/app/oracle/product/19.0.0.0/dbhome_1
Perpatchoperation completed successfully on home/app/app/oracle/product/19.0.0.0/dbhome_1

Startapplying binary patch on home /app/app/oracle/product/19.0.0.0/dbhome_1
Failedwhile applying binary patches on home /app/app/oracle/product/19.0.0.0/dbhome_1

Executionof [OPatchAutoBinaryAction] patch action failed, check log for more details.Failures:
PatchTarget : testrac1->/app/app/oracle/product/19.0.0.0/dbhome_1 Type[rac]
Details:[
---------------------------PatchingFailed---------------------------------
Commandexecution failed during patching in home:/app/app/oracle/product/19.0.0.0/dbhome_1, host: testrac1.
Commandfailed: /app/app/oracle/product/19.0.0.0/dbhome_1/OPatch/opatchauto  apply /app/soft/32226239 -oh/app/app/oracle/product/19.0.0.0/dbhome_1 -target_type rac_database -binary-invPtrLoc /app/app/19.0.0.0/grid/oraInst.loc -jre/app/app/19.0.0.0/grid/OPatch/jre -persistresult/app/app/oracle/product/19.0.0.0/dbhome_1/opatchautocfg/db/sessioninfo/sessionresult_testrac1_rac_2.ser-analyzedresult /app/app/oracle/product/19.0.0.0/dbhome_1/opatchautocfg/db/sessioninfo/sessionresult_analyze_testrac1_rac_2.ser
Commandfailure output: 
==Followingpatches FAILED in apply:

Patch:/app/soft/32226239/32218454
Log:/app/app/oracle/product/19.0.0.0/dbhome_1/cfgtoollogs/opatchauto/core/opatch/opatch2022-03-04_15-30-20PM_1.log
Reason: Failed during Patching:oracle.opatch.opatchsdk.OPatchException: Prerequisite check"CheckActiveFilesAndExecutables" failed. 

Afterfixing the cause of failure Run opatchauto resume]
OPATCHAUTO-68061:The orchestration engine failed.
OPATCHAUTO-68061:The orchestration engine failed with return code 1
OPATCHAUTO-68061:Check the log for more details.
OPatchAutofailed.
OPatchautosession completed at Thu Mar  4 15:31:152022
Timetaken to complete the session 3 minutes, 25 seconds

 opatchauto failed with error code 42

查看日志

tail -100f/app/app/oracle/product/19.0.0.0/dbhome_1/cfgtoollogs/opatchauto/core/opatch/opatch2022-03-04_15-30-20PM_1.log

[Mar4, 2022 3:31:14 PM] [INFO]     Finishfuser command /bin/fuser /app/app/oracle/product/19.0.0.0/dbhome_1/bin/expdp atThu Mar 04 15:31:14 CST 2021
[Mar4, 2022 3:31:14 PM] [INFO]    Followingactive executables are not used by opatch process :                                /app/app/oracle/product/19.0.0.0/dbhome_1/lib/libclntsh.so.19.1                                 /app/app/oracle/product/19.0.0.0/dbhome_1/lib/libsqlplus.so                              
Followingactive executables are used by opatch process :
[Mar4, 2022 3:31:14 PM] [INFO]    Prerequisite check "CheckActiveFilesAndExecutables" failed.
                                    The detailsare:

 Following active executables are not used byopatch process :
                                    /app/app/oracle/product/19.0.0.0/dbhome_1/lib/libclntsh.so.19.1                               /app/app/oracle/product/19.0.0.0/dbhome_1/lib/libsqlplus.so
  Following active executables are used byopatch process :
[Mar4, 2022 3:31:15 PM] [SEVERE]   OUI-67073:UtilSessionfailed: Prerequisite check "CheckActiveFilesAndExecutables" failed.
[Mar4, 2022 3:31:15 PM] [INFO]     FinishingUtilSession at Thu Mar 04 15:31:15 CST 2022
[Mar4, 2021 3:31:15 PM] [INFO]     Log filelocation: /app/app/oracle/product/19.0.0.0/dbhome_1/cfgtoollogs/opatchauto/core/opatch/opatch2022-03-04_15-30-20PM_1.log

查看进程占用情况

fuser/app/app/oracle/product/19.0.0.0/dbhome_1/lib/libclntsh.so.19.1
/app/app/oracle/product/19.0.0.0/dbhome_1/lib/libclntsh.so.19.1:89500m 109054m

kill 掉占用的进程

kill-9 89500

kill-9 109054

验证没有进程占用

fuser/app/app/oracle/product/19.0.0.0/dbhome_1/lib/libsqlplus.so
fuser /app/app/oracle/product/19.0.0.0/dbhome_1/lib/libclntsh.so.19.1

继续打补丁

testrac1:~# /app/app/19.0.0.0/grid/OPatch/opatchauto resume

4、CRS-6706 patchlevel 不一致

19.3.RAC 的 2 节点打补丁失败，导致2节点集群无法启动报错：

[root@testrac2 soft]#/u01/app/19.0.0.0/grid/bin/crsctl start crs
CRS-6706: OracleClusterware Release patch level ('4203896349') does not match Software patchlevel ('724960844'). Oracle Clusterware cannot be started.
CRS-4000: Command Startfailed, or completed with errors.

从报错补丁 patch level 不一致导致。然后查看 mos

1.  Run the following command as the root user tocomplete the patching set up behind the scenes:

#GI_HOME/bin:>  ./clscfg -localpatch

2.  Run the following command as the root user tolock the GI home:

#GI_HOME/crs/install:>  ./rootcrs.sh -lock

3.  Run the following command as the root user tostart the GI:

#GI_HOME/bin:>  ./crsctl start crs

执行：
[root@testrac2 bin]#./clscfg -localpatch
[root@testrac2install]# ./rootcrs.sh -lock
[root@testrac2 bin]#./crsctl start crs

然后集群可以启动。

参考文档：

CRS-6706: OracleClusterware Release patch level ('nnn') does not match Software patch level('mmm') (文档 ID 1639285.1)

Patching 12.2.0.1 GridInfrastructure gives error CRS-6706: Oracle Clusterware Release Patch Level('748994161') Does Not Match Software Patch Level (文档 ID 2348013.1)