Oracle Clusterware and Oracle Real Application Clusters
Installation Guide
10g Release 2 (10.2) for Solaris10 x86-64 Operating System
#In this article I'll describe the installation of Oracle Clusterware and Oracle Real Application Clusters on Solaris 10 (x86-64) on vmware workstation 8.0.
#
#To determine whether the system architecture can run the Oracle software you have obtained:
/bin/isainfo -kv
#64-bit amd64 kernel modules
#Software
#10201_clusterware_solx86_64
#10201_database_solx86_64
#pre-installation
#1.Checking the Hardware Requirements
#Each system must meet the following minimum hardware requirements
#At least 1 GB of physical RAM
#Swap space equivalent to the multiple of the available RAM, as the following.
#Between 1 GB and 2 GB 1.5 times the size of RAM
#More than 2 GB Equal to the size of RAM
# note:
/usr/sbin/prtconf | grep "Memory size"--To determine the physical RAM size
/usr/sbin/swap -s ---to determine the size of configured swap space
how to add or reduce swap size:
eg:swap -a /dev/dsk/c1t1d0s3#temporary# append an entry into /etc/vfstab
#permanent#
/dev/dsk/c1t5d0s3 - - swap - no -
swap -d /dev/dsk/c1t1d0s3
verify the size of swap space #df -k /tmp
#mount disks for ocr and vote.at least 120m for ocr,20m for vote disk.
#note:
# 1. Determine which disk devices will be used store the OCR and voting disk file. Use the format
# command to list the devices, and select the appropriate device. Create a 120 Mb partition and
# label the disk. Note, the disk partition to be used must skip the first 1Mb to avoid overwriting
# the disk VTOC.
# Repeat this step for the voting disk, using 20Mb for the disk device.
# 2. Make sure the other nodes can see those devices. If the other nodes see those devices at a
# different path, then common symbolic links or mknod command must be used to point to the same location.
# 3.normally,i don't use sector 0 to create partition
#Change the permissions and ownership on the two devices.
#OCR disk device
chown root:dba /dev/rdsk/c1t1d0s*
chmod 660 /dev/rdsk/c1t1d0s3
#Voting disk
chown oracle:dba /dev/rdsk/c1t2d0s*
chmod 660 /dev/rdsk/c1t2d0s3
#2.Checking the Network Requirements
#check the following instances:
#edit /etc/hosts
::1 localhost
127.0.0.1 localhost
192.168.58.2 dms01 loghost
192.168.58.3 dms02
#priv
192.168.117.20 dms01-priv
192.168.117.30 dms02-priv
#vip
192.168.58.5 dms01-vip
192.168.58.6 dms02-vip
#create &edit /etc/defaultrouter
#input gateway address#if you will be missing this step,you will get an error like " Default gateway is not defined" or" No more members to consider"
#we mush reach getway address,if not ,vip will be crash and swith over.
#change hostname
nodename
#reboot.
#To determine User equivalence for all node.
#perform. the following action on one node.--here,it is node A,dms01
#log in as oracle
mkdir .ssh
chmod 700 .ssh
cd .ssh
ssh-keygen -t rsa
ssh-keygen -t dsa
touch authorized_keys
ssh dms02 cat /dms01/app/.ssh/*pub >>authorized_keys
ssh dms01 cat /dms01/app/.ssh/*pub >>authorized_keys
#copy to
scp authorized_keys dms02:/dms01/app/.ssh
#verify the equivalence on all node simply.
ssh dms01 date
ssh dms02 date---if prompt input password,there would be a wrong for this step.
#verify by cluvfy.
./runcluvfy.sh stage -pre crsinst -n dms01,dms02 -verbose
#normally,it will raise following error here.
#Result: User equivalence check failed for user "oracle".
#
#ERROR:
#User equivalence unavailable on all the nodes.
#Verification cannot proceed.
#solution
mkdir -p /usr/local/bin
ln -s -f /usr/bin/ssh /usr/local/bin/ssh
ln -s -f /usr/bin/scp /usr/local/bin/scp
exec /usr/bin/ssh-agent $SHELL
/usr/bin/ssh-add
#verify once more.
#3.Identifying Software Requirements
pkginfo -i SUNWarc SUNWbtool SUNWhea SUNWlibm SUNWlibms SUNWsprot SUNWsprox SUNWtoo SUNWi1of SUNWi1cs SUNWi15cs SUNWxwfnt
#i missed three packages.
#ERROR: information for "SUNWsprox" was not found
#ERROR: information for "SUNWi1cs" was not found
#ERROR: information for "SUNWi15cs" was not found
#mount cdrom and then install package like:
pkgadd -d /cdrom/sol_10_910_x86/Solaris_10/Product SUNWi1cs
#note:As of Solaris 10 the packages are 64-bit and the x-packages moved: SUNWsprox is SUNWsprot.
#4.Configuring Kernel Parameters
#edit /etc/system add entries as below
set noexec_user_stack=1
set semsys:seminfo_semmni=100
set semsys:seminfo_semmns=1024
set semsys:seminfo_semmsl=256
set semsys:seminfo_semvmx=32767
set shmsys:shminfo_shmmax=4294967296
set shmsys:shminfo_shmmni =100
#note:The Oracle installer recognizes kernel parameters set using this method, but it is now deprecated in favour of resource control projects, explained below.
#we also can use following method
projadd -U oracle -K "project.max-shm-memory=(priv,4096MB,deny)" oracle
#Append the following line to the "/etc/user_attr" file.
oracle::::project=oracle
#verify after reboot.
prctl -n project.max-shm-memory -i project oracle
#5.Create the new groups and users.
groupadd oinstall
groupadd dba
groupadd oper--optional
useradd -g oinstall -G dba[oper] -d /dms01/app oracle
#6.set environment
$ cat .profile
umask 022
ORACLE_BASE=/dms01/app
ORACLE_SID=dms01
ORACLE_HOME=$ORACLE_BASE/oracle
CRS_ORACLE_HOME=$ORACLE_BASE/crs
PATH=$ORACLE_HOME/bin:$CRS_ORACLE_HOME/bin:$PATH
TEMP=/tmp
TMPDIR=$TEMP
export ORACLE_BASE ORACLE_SID ORACLE_HOME CRS_ORACLE_HOME PATH TEMP TMPDIR
mkdir -p $ORACLE_HOME
mkdir -p $CRS_ORACLE_HOME
#7.Installation
#log in as the oracle user on server directly.or switch to oracle from root .
#root
xhost +
#swith to oracle
DISPLAY=:0.0; export DISPLAY
#Start the Oracle Universal Installer (OUI) by issuing the following command in the database directory.
./runInstaller
the other issue.
1.unable to log on to server by root.
cd /ect/ssh
edit sshd_config
enable PasswordAuthentication --PasswordAuthentication yes
enable PermitRootLogin --PermitRootLogin yes
reboot.
2.missing network card.
ifconfig -a --check currently network card
ifconfig -a plumb --plumb new card.
touch new file 'hostname.e1000g1' under /etc
edit /add ip to 'hostname.e1000g1'
or make it available temporary.--example :ifconfig e1000g1 10.0.0.2 netmask 255.0.0.0 up
edit hosts under hosts.
reboot.
note--------check node name// /etc/nodename more configuration under /etc/inet
3.fail to install crs
clean all file relate to crs
delete crs home
delete /var/opt/oracle/ocr.loc
clean ocr disk
dd if=/dev/zero f=/dev/rdsk/c1t1d0s3--ocr
dd if=/dev/zero f=/dev/rdsk/c1t2d0s3--voting disk
mv -f /etc/init.d/init.cssd /etc/init.d/init.cssd.bak
mv -f /etc/init.d/init.crs /etc/init.d/init.crs.bak
mv -f /etc/init.d/init.crsd /etc/init.d/init.crsd.bak
mv -f /etc/init.d/init.evmd /etc/init.d/init.evmd.bak
mv -f /etc/inittab.crs /etc/inittab.crs.bak
rm -f /etc/rc2.d/K96init.crs
rm -f /etc/rc2.d/S96init.crs
rm -f /etc/rc3.d/K96init.crs
rm -f /etc/rc3.d/S96init.crs
rm -f /etc/rc5.d/K96init.crs
rm -f /etc/rc5.d/S96init.crs
5.we alway get the following error when we ran root.sh on first node.it could be ignore.
2012-04-21 23:03:03.931
[client(2352)]CRS-1006:The OCR location /dev/rdsk/c1t1d0s3 is inaccessible. Deta
ils in /dms01/app/crs/log/dms01/client/ocrconfig_2352.log.
2012-04-21 23:03:04.036
[client(2352)]CRS-1006:The OCR location /dev/rdsk/c1t1d0s3 is inaccessible. Deta
ils in /dms01/app/crs/log/dms01/client/ocrconfig_2352.log.
2012-04-21 23:03:04.038
[client(2352)]CRS-1006:The OCR location /dev/rdsk/c1t1d0s3 is inaccessible. Deta
ils in /dms01/app/crs/log/dms01/client/ocrconfig_2352.log.
2012-04-21 23:03:04.351
[client(2352)]CRS-1001:The OCR was formatted using version 2.
7.share disk
disk.locking = "FALSE"
来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/21993926/viewspace-722065/,如需转载,请注明出处,否则将追究法律责任。
转载于:http://blog.itpub.net/21993926/viewspace-722065/