AIX-6.1下RAC安裝問題小結(jié)

.@;AIX6.1下RAC安裝問題小結(jié)AIX6.1,RAC分享到：新浪微博騰訊微博QQ空間QQ好友更多0最近在AIX6.1上安裝了一套R(shí)AC，實(shí)施得不多，遇到不少問題，記錄一下：

OS版本：

HOST_NAM_1:/#oslevel-s

6100-04-02-1007

HA版本：

HOST_NAM_1:/#lslpp-lcluster.*

Fileset

Level

State

Description

----------------------------------------------------------------------------

Path:/usr/lib/objrepos

cluster.adt.es.client.include

COMMITTED

ESClientIncludeFiles

cluster.adt.es.client.samples.clinfo

COMMITTED

ESClientCLINFOSamples

cluster.adt.es.client.samples.clstat

COMMITTED

ESClientClstatSamples

###RSH報(bào)錯(cuò)

#rshHOST_NAM_2date

rshd:0826-813Permissionisdenied.

相關(guān)文件配置：

#cat.rhosts

HOST_NAM_1root

HOST_NAM_2root

HOST_NAM_1oracle

HOST_NAM_2oracle

#cat/etc/hosts.equiv

HOST_NAM_1root

HOST_NAM_2root

HOST_NAM_1oracle

HOST_NAM_2oracle

其中HOST_NAM_1、HOST_NAM_2是HOSTNAME。

這里主要是/etc/hosts文件中，HOSTNAME不能當(dāng)作別名，或者，”.rhosts”、”hosts.equiv”里不要配別名，應(yīng)該是跟解析有關(guān)。

原HOSTS配置：

1

HOST_NAM_1_boot1

HOST_NAM_1

2

HOST_NAM_2_boot1

HOST_NAM_2

7

HOST_NAM_1_boot2

8

HOST_NAM_2_boot2

6

HOST_NAM_2_vip

5

HOST_NAM_1_vip

改為：

1

HOST_NAM_1_boot1

2

HOST_NAM_2_boot1

7

HOST_NAM_1

HOST_NAM_1_boot2

8

HOST_NAM_2

HOST_NAM_2_boot2

6

HOST_NAM_2_vip

5

HOST_NAM_1_vip

###rootpre.sh報(bào)錯(cuò)

這個(gè)安裝前在文檔中有到看到，作為注意事項(xiàng)記錄一下：

TheOracle10gR2OUIandconfigurationassistantprogramsdonotrecognizeAIX6V6.1asasupportedrelease.

執(zhí)行rootpre.sh時(shí)會(huì)報(bào)：

ConfiguringAsynchronousI/O....

AsynchronousI/Oisnotinstalledonthissystem.

Youwillneedtoinstallit,andeitherconfigureityourselfusing

'smitaio'orreruntheOraclerootinstallationprocedure.

ConfiguringPOSIXAsynchronousI/O....

PosixAsynchronousI/Oisnotinstalledonthissystem.

Youwillneedtoinstallit,andeitherconfigureityourselfusing

'smitaio'orreruntheOraclerootinstallationprocedure.

解決方法：下載6718715補(bǔ)丁，執(zhí)行里面的rootpre.sh

參考文檔：282036.1

###VIPCA報(bào)錯(cuò)

VIPCA時(shí)，VIP起不來，日志報(bào)錯(cuò)信息：

Interfaceen4checkedfailed(host=HOST_NAM_1)

Invalidparameters,orfailedtobringupVIP(host=HOST_NAM_1)

原因：VIP綁定的是小機(jī)集成的網(wǎng)卡LogicalHostEthernetPort(lp-hea)

TheentstatoutputforLHEAisdifferentfromaregularadapter

解決方法：

修改racgvip腳本，找到

$ENTSTAT-d$_IF這行，修改為：

$ENTSTAT-d$_IF|$GREP-iEq'.*lan.*state.*:.*operational.*|.*link.*status.*:.*up.*|.*port.*operational.*state.*:.*up.*|.*driver.*flags.*:.*up.*'

參考文檔：959746.1

###ONS起不來

日志報(bào)錯(cuò)信息：

FailedtogetIPforlocalhost(0)

FailedtogetIPforlocalhost(0)

FailedtogetIPforlocalhost(0)

onsctl:onsfailedtostart

解決方法：

原h(huán)osts文件中找不到localhost：

loopback

改為：

loopbacklocalhost

###CRS升級(jí)報(bào)錯(cuò)

升級(jí)完成后執(zhí)行root102.sh，報(bào)：

#./root102.sh

Error:PleasechangetheCRS_ORACLE_USERidoracle

tohavethefollowingOScapabilities:

<CAP_PROPAGATECAP_BYPASS_RAC_VMMCAP_NUMA_ATTACH>

解決方法：

#chusercapabilities=CAP_BYPASS_RAC_VMM,CAP_PROPAGATE,CAP_NUMA_ATTACHoracle

#lsuser-foracle|grepcapabilities

capabilities=CAP_BYPASS_RAC_VMM,CAP_PROPAGATE,CAP_NUMA_ATTACH

這個(gè)報(bào)錯(cuò)之前遇到過，升級(jí)的文檔中，也有提到。

###CRS升級(jí)后，VIP起不來

這次的有點(diǎn)難搞，日志中沒有太多的信息，只有一行：

Invalidparameters,orfailedtobringupVIP(host=HOST_NAM_1)

后來使用crsctl對(duì)VIP進(jìn)行debug，收集更多的信息：

#crsctldebuglogres"ora.host_nam_2.vip:5"

SetResourceDebugModule:ora.host_nam_2.vip

Level:5

#srvctlstartnodeapps-nhost_nam_2

CRS-0233:Resourceorrelativesarecurrentlyinvolvedwithanotheroperation.

host_nam_2ra.host_nam_2.vip:ThuMar2513:57:25GMT+08:002010[360824]Checkinginterfaceexistance

host_nam_2ra.host_nam_2.vip:ThuMar2513:57:25GMT+08:002010[360824]Callinggetifbyip

host_nam_2ra.host_nam_2.vip:ThuMar2513:57:25GMT+08:002010[360824]getifbyip:

startedfor6

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:26GMT+08:002010[360824]getifbyip:checkingiffailoverishappening()

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:26GMT+08:002010[360824]getifbyip:failoverisnothappening()

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:26GMT+08:002010[360824]Completedgetifbyip

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:26GMT+08:002010[360824]ping_vip6started

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:26GMT+08:002010[360824]Abouttoexecute:/usr/sbin/ping-c1-w16

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]ping_vip:6isnotpingable,_count=1

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]Completedwithinitialinterfacetest

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]Broadcast=55

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]Interfacetests

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]checkIf:startforif=en4

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]IsIfAlive:startforif=en4

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]defaultgw:

started

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]defaultgw:

completedwith54

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:27GMT+08:002010[360824]Abouttoexecutecommand:/usr/sbin/ping-S8-c1-w154

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:28GMT+08:002010[360824]Abouttoexecutecommand:/usr/sbin/ping-S8-c1-w154

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:29GMT+08:002010[360824]IsIfAlive:RXpacketscheckedif=en4failed

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:29GMT+08:002010[360824]Interfaceen4checkedfailed(host=HOST_NAM_2)

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:29GMT+08:002010[360824]IsIfAlive:endforif=en4

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:29GMT+08:002010[360824]checkIf:endforif=en4

host_nam_2:ora.host_nam_2.vip:Invalidparameters,orfailedtobringupVIP(host=HOST_NAM_2)

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:30GMT+08:002010[307376]Checkinginterfaceexistance

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:30GMT+08:002010[307376]Callinggetifbyip

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:30GMT+08:002010[307376]getifbyip:

startedfor6

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:30GMT+08:002010[307376]getifbyip:checkingiffailoverishappening()

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:30GMT+08:002010[307376]getifbyip:failoverisnothappening()

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:30GMT+08:002010[307376]Completedgetifbyip

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:30GMT+08:002010[307376]ping_vip6started

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:30GMT+08:002010[307376]Abouttoexecute:/usr/sbin/ping-c1-w16

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:31GMT+08:002010[307376]ping_vip:6isnotpingable,_count=1

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:31GMT+08:002010[307376]Completedwithinitialinterfacetest

host_nam_2:ora.host_nam_2.vip:ThuMar2513:57:31GMT+08:002010[307376]Broadcast=55

CRS-1006:Nomorememberstoconsider

CRS-0215:Couldnotstartresource'ora.host_nam_2.vip'.

CRS-0210:Couldnotfindresourceora.host_nam_2.LISTENER_HOST_NAM_2.lsnr.

這才搜索到了相關(guān)信息：

Bug8413088:VIPCANNOTSTARTONAIX6.1BECAUSENETSTATHASANEWCOLUMN.

Bug9157855:DURINGRESTARTORWHENONEOFTHETWONODECLUSTERISDOWN,VIPRESOURCEFAILS

這個(gè)問題，在打完CRSPSU后，也同樣有可能存在，可以通過修改racvip腳本來解決：

：

_O1=`$NETSTAT-n-I$_IF|$AWK"{if(/^$_IF/){print\\$5;exit}}"`

_O2=`$NETSTAT-n-I$_IF|$AWK"{if(/^$_IF/){print\\$5;exit}}"`

打完P(guān)SU后：

_O1=`$NETSTAT-n-I$_IF-ptcp|$GREP-iE".*packetsreceived$"|$AWK"{print\\$1;exit}"`

_O2=`$NETSTAT-n-I$_IF-ptcp|$GREP-iE".*packetsreceived$"|$AWK"{print\\$1;exit}"`

最終改為：

_O1=`$NETSTAT-n-I$_IF-pip|$GREP-iE".*packetsreceived$"|$AWK"{print\\$1;exit}"`

_O2=`$NETSTAT-n-I$_IF-pip|$GREP-iE".*packetsreceived$"|$AWK"{print\\$1;exit}"`

這個(gè)問題很杯具，花了很多時(shí)間，安裝前，閱讀相關(guān)文檔時(shí)，我就注意到了這個(gè)BUG，兩次VIP起不來，我都拿去對(duì)比，看看是不是這個(gè)BUG。結(jié)果還是沒發(fā)現(xiàn)，一直到debug出來。

###RDBMS升10204時(shí)，報(bào)java進(jìn)程沒停

OracleUniversalInstallerhasdetectedthat

thereareprocessesrunninginthe

currentlyselectedOracleHome.The

followingprocessesneedtobeshutdown

beforecontinuing:

java

剛開始時(shí)，還有CRS的進(jìn)程，停掉CRS后，還有一個(gè)java始終過不去，用fuser查使用$ORACLE_HOME目錄進(jìn)程，全kill，把ps-ef|grepjava出來的進(jìn)程，除了安裝的進(jìn)程外都?xì)⒘?，還是不行。

最后在Metalink上找到了解決方法，升級(jí)前：

cd/usr/sbin/

mvfuserfuser.orig

touch/usr/sbin/fuser

chmod+x/usr/sbin/fuser

升級(jí)完成后，再改回來：

cd/usr/sbin/

cpfuser.origfuser

這招很陰啊。。

參考文檔：975597.1

###數(shù)據(jù)庫(kù)打完P(guān)SU補(bǔ)丁后，啟CRS報(bào)錯(cuò)

在CRS、database都升級(jí)、打補(bǔ)丁完成后，啟CRS、VIP等資源時(shí)，報(bào)錯(cuò)：

HOST_NAM_1:/#crsctlstartcrs

exec():0509-036Cannotloadprogram/app/oracle/product/10204/db_1/bin/crsctl.binbecauseofthefollowingerrors:

0509-150Dependentmodulelibhasgen10.a(shr_hasgen10.o)couldnotbeloaded.

0509-022Cannotloadmodulelibhasgen10.a(shr_hasgen10.o).

0509-026Systemerror:Afileordirectoryinthepa