在AIX平台安装Oracle RAC就是个坑爹的事情,今日再次验证了这点,只因为bug太多!记录问题也只因为bug太多!强烈推荐及看好Oracle RAC在Linux平台的发展。下面对问题的记录:
>服务器:IBM Power 740
>存储:IBM DS 5020
> 操作系统:
mzrac2@root[/oracle]oslevel -s
6100-07-00-0000
> Oracle Grid&Database:11.2.0.3
一.安装过程出现如下问题:
1.在检查过程中出现如下的告警:
经过检查,在AIX6.1TL7上是不需要这两个补丁的,因此可以直接忽略。
2.安装Grid的时候出现如下的报错:
点击"Retry"之后又可以顺利通过。
之后安装Database软件的时候比较顺利。
3.使用dbca创建数据库的时候问题又来了,创建过程中始终报ORA-03113的错误,如下图:
虽然报错,但是数据库又能继续创建,直到创建成功!坑爹啊,这么创建的数据库谁敢用啊!!!
4.core_* dump目录及文件撑爆你的$ORACLE_HOME/dbs目录。
在这种情况下创建的数据库会在$ORACLE_HOME/dbs目录下生成大量的core_* dump目录及文件,直到文件系统被撑爆!!!
--------------------------------------------------------------------------------
在CentOS 6.4下安装Oracle 11gR2(x64)
--------------------------------------------------------------------------------
二.下面是排错记录:
1.在METALINK上找到如下的文章说明了ORA-03113的错误:
"ld: 0711-780 SEVERE ERROR: Symbol .ksmpclrpga" and "ORA-7445 [KSMPCLRPGA()+23248]" [ID 1379908.1]
修改时间 27-JUN-2012 类型 PROBLEM 状态 PUBLISHED
In this Document
Oracle Server - Enterprise Edition - Version 11.2.0.1 to 11.2.0.3 [Release 11.2]
IBM AIX on POWER Systems (64-bit)
Installation / Relink of 11.2.0.x on AIX with following version
AIX 5.3 TL 8 and above
AIX 6.1 TL 6 and above
AIX 7.1 SP3 (7100.00.03.1115), TL 1 (7100-01-01-1141) and above
Fail with following relink errors :
ld: 0711-783 WARNING: TOC overflow. TOC size: 225720 Maximum size: 65536
Extra instructions are being generated for each reference to a TOC
symbol if the symbol is in the TOC overflow area.
ld: 0711-780 SEVERE ERROR: Symbol .ksmpfpva (entry 58964) in object /oracle/lib//libserver11.a[ksmp.o]:
Relocation overflow in reference to: dbkea_prm_rate_ (entry 8601)
Address: 0x00008056; RLD type: R_TOC; RLD length: 16
ld: 0711-780 SEVERE ERROR: Symbol .ksmpfpva (entry 58964) in object /oracle/lib//libserver11.a[ksmp.o]:
Relocation overflow in reference to: dbkea_prm_rtl_ (entry 8603)
Address: 0x0000805e; RLD type: R_TOC; RLD length: 16
ld: 0711-780 SEVERE ERROR: Symbol .ksmpfpva (entry 58964) in object /oracle/lib//libserver11.a[ksmp.o]:
Relocation overflow in reference to: dbkea_prm_ctl_ (entry 8605)
Address: 0x00008062; RLD type: R_TOC; RLD length: 16
ld: 0711-780 SEVERE ERROR: Symbol .ksmpfpva (entry 58964) in object /oracle/lib//libserver11.a[ksmp.o]:
Relocation overflow in reference to: dbkehackflg_ (entry 8607)
Address: 0x0000806a; RLD type: R_TOC; RLD length: 16
ld: 0711-780 SEVERE ERROR: Symbol .ksmpfpva (entry 58964) in object /oracle/lib//libserver11.a[ksmp.o]:
Relocation overflow in reference to: dbkedPix_ (entry 8609)
Address: 0x00008076; RLD type: R_TOC; RLD length: 16
ld: 0711-780 SEVERE ERROR: Symbol .ksmpfpva (entry 58964) in object /oracle/lib//libserver11.a[ksmp.o]:
Relocation overflow in reference to: dbkehackprm_ (entry 8611)
Address: 0x0000807e; RLD type: R_TOC; RLD length: 16
ld: 0711-780 SEVERE ERROR: Symbol .ksmpfpva (entry 58964) in object /oracle/lib//libserver11.a[ksmp.o]:
Relocation overflow in reference to: dbkeddmper_ (entry 8613)
Address: 0x00008082; RLD type: R_TOC; RLD length: 16
Further, if the above relink error is ignored while installation and database is created OR upgraded, then following error is noticed while shutting down the database.
ORA-03113: end-of-file on communication channel
ChangesAIX O.S. was installed / upgraded to a new TL level
CauseThis is due to the
"Bug 13342637: 0711-780 SEVERE ERROR: SYMBOL .KSMPFPVA AND .KSMPCLRPGA"
Other bugs referenced / raised for similar issue
Unpublished Bug 10380509: ORACLE LINK ISSUE
Bug 13354044: LINK ERROR - LD: 0711-780 SYMBOL .KSMPFPVA IN OBJECT LIBSERVER11.A[KSMP.O]
BUG 13384572 - ORA-7445 [KSMPCLRPGA()+23248]
Please work with IBM to get a fix for your OS Version, as you look to be running into AIX O.S. issue, you need to provide your current O.S. level to IBM to get the appropriate fix.
Following are the iFixes need to be applied
- Note that you need to download the right ifix for your OS version and Technology level
For 5.3 -- IV10538 FILE.ATION OVERFLOW REPORTED IN ERROR WHILE LINKING LARGE" (for v5.3)
https://www-304.ibm.com/support/docview.wss?uid=isg1IV10538
For AIX 6.1 -- IV09580: FILE.ATION OVERFLOW REPORTED IN ERROR WHILE LINKING LARGE (for v6.1)
https://www-304.ibm.com/support/docview.wss?uid=isg1IV09580
For 7.1 IFIX IV09541 -- IV09541: FILE.ATION OVERFLOW REPORTED IN ERROR WHILE LINKING LARGE (for v7.1)
https://www-304.ibm.com/support/docview.wss?uid=isg1IV09541
Below are the locations of the prebuilt iFixes.
ftp://public.dhe.ibm.com/aix/efixes/<apar#>/
where the APAR numbers are:
5.3 TL11 - iv10538
5.3 TL12 - iv11158
6.1 TL4 - iv11167
6.1 TL5 - iv10576
6.1 TL6 - iv10539
6.1 TL7 - iv09580
7.1 TL0 - unaffected
7.1 TL1 - iv09541
For example:
ftp://public.dhe.ibm.com/aix/efixes/iv09541/
Note : Confirm with IBM, before applying the fixes
After installing these fixes, execute "relink all" again, verify that there is no "ld: 0711-780" in the log. You may get some "ld" warnings
ld: 0711-224
ld: 0711-345
ld: 0711-773
ld: 0711-783
ld: 0711-301
ld: 0711-415
ld: 0711-319
These warnings can be safely ignored, refer to following documents
Note 1189533.1 Relinking causes many warning on AIX
Note 245372.1 TOC overflow Warning Can Safely Be Ignored
Note 809153.1 "Relink All" On AIX Raises Many Warnings And "ld: 0706-032 Linking mode is not specified" for Liborasdk
从上面的内容不难看出,出现ORA-03113的错误是由于AIX操作系统的bug所致,需要将操作系统升级到AIX 6.1.7.1的版本,或者应用紧急补丁iv09580。
iv09580紧急补丁的描述:https://www-304.ibm.com/support/docview.wss?uid=isg1IV09580
iv09580紧急补丁的下载:ftp://public.dhe.ibm.com/aix/efixes/iv09580
2.使用操作系统的emgr命令应用iv09580补丁 :
从上面的地址下载iv09580紧急补丁,执行下面的步骤应用紧急补丁。
1).紧急补丁安装预览命令:
#emgr -p -e IV09580.epkg.Z
出现INSTALL PREVIEW ,SUCCESS的情况下才能执行后面的安装命令。
2).应用紧急补丁:
#emgr -e IV09580.epkg.Z
3).查看补丁情况:
mzrac1@root[/]emgr -l
ID STATE LABEL INSTALL TIME UPDATED BY ABSTRACT
=== ===== ========== ================= ========== ======================================
1 S IV09580s01 06/27/12 21:55:58 Ifix for IV09580@6.1TL7SP1
STATE codes:
S = STABLE
M = MOUNTED
U = UNMOUNTED
Q = REBOOT REQUIRED
B = BROKEN
I = INSTALLING
R = REMOVING
T = TESTED
P = PATCHED
N = NOT PATCHED
SP = STABLE + PATCHED
SN = STABLE + NOT PATCHED
QP = BOOT IMAGE MODIFIED + PATCHED
QN = BOOT IMAGE MODIFIED + NOT PATCHED
RQ = REMOVING + REBOOT REQUIRED
应用了此补丁之后,需要重新安装Grid和Database,之后再创建数据库。在重新安装Grid的时候再没出现编译ins_rdbms.mk文件的问题,且后来创建数据库也非常的顺利,至此问题得到最终解决!
在整个问题解决的过程中还参考了另外两篇文章,是在AIX7.1上安装11.2.0.3遇到跟上面描述几乎完全相同的问题。