Print it. Fill in during installation.
IP addresses.
|
Public
name |
Public
IP |
Private
name |
Private
IP |
VIP
name |
VIP
IP |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Script |
NODE1 |
NODE2 |
Comment |
|
001-patch-iscsi.sh |
|
|
optional |
|
005-copy-rawnames.sh |
|
|
|
|
009-now-reboot-everything.sh |
|
|
optional |
|
Make sanity check 1 here, |
|||
|
010-install-orarun.sh |
|
|
|
|
015-edit-oraprofile.sh |
|
|
|
|
018-edit-rcoracle.sh |
|
|
|
|
020-start-rcoracle.sh |
|
|
|
|
030-check-oracle-user.sh |
|
|
|
|
060-setup-xntp.sh |
|
|
optional |
|
070-edit-services.sh |
|
|
(not required?) |
|
080-ssh-genkeys.sh |
|
|
|
|
090-verify-ssh-access.sh |
|
|
|
|
Make sanity check
here. Verify X11 access. |
|||
|
100-verify-rawfiles.sh |
|
|
|
|
200-InstCRS.1sh |
|
*** |
1 node only |
|
210-runInventory.sh |
|
|
Sometimes skipped |
|
220-runRootsh.sh |
|
|
|
|
230-check-crs.sh |
|
|
|
|
300-check-env.sh |
|
|
|
|
310-InstOracle.1sh |
|
*** |
1 node only |
|
315-Apply-patch4045013.sh |
|
|
x86_64 only |
|
320-RunRootSh.sh |
|
|
|
|
340-run-netca.1sh |
|
***
|
1 node only |
|
350-relink-with-aio.sh |
|
|
optional |
|
400-create-database.1sh |
|
|
|
|
410-edit-oratab.sh |
|
|
|
|
Make sanity check here. Verify installation, start and stop sequences. |
|||
1. System and network configuration. Run it after you completed iSCSI, RAW and network configuration steps.
a. Public IP allocated, resolved and configured;
b. Private IP allocated, resolved (as node-1) and configured;
c. You use 1Gbit interface for PRIVATE IP;
d. Duplex and SPEED are thee same on server and on Ethernet switch;
e. VIP IP (in PUBLCI network) are allocated and resolved.
f.
No wrong records in /etc/hosts (often created by
installation). If you use /etc/hosts, and node name is rac, you must see
something like this in /etc/hosts:
10.23.32.11 rac1 rac1.domain.com
10.24.32.11 rac1-1
10.23.32.12 rac1-vip
10.23.32.13 rac2 rac2.domain.com
10.24.32.13 rac2-1
10.24.32.14 rac2-vip
…
ERRORS here are primary reason
for 95% of installation problem, so DOUBLE CHECK.
g. You can ping rac1, rac2, rac1-1, rac2-1 from BOTH nodes (replace rac by your host name);
h. You have enough swap (2Gb);
i. You have enough space in /tmp;
j. You have enough space in /opt.
2. ISCSI configuration:
a. ISCSI id are configured, are different OR are generated;
b. iscsi is enabled as a service;
c. iscsi-ls –l shows all disks;
d. ls /dev/disk/by-id shows all disk ID’s;
3. RAW device configuration (rawnames.conf and script rawnames):
a. You configured raw devices for CSSFile (25Mb), OCRFile (100Mb) and for ASM (at least 4 Gb);
b. You can read (od /dev/rawnames/NAME) all 3 of this files; DO NOT SKIP THIS STEP;
c. /etc/rawnames/* files are available on both nodes.
4. Reboot.
a. You can reboot system (try on all nodes) and everything comes in order (iscsi, rawnames, IP addresses, duplex mode).
5. If you use installation scripts, check that you configured CONFIG.sh:
a. node names are correct (for both, public and private, names);
b. SID names are correct for your test database;
c. SID names are different for different nodes (test1 for node1 and test2 for node2, for example);
d. You did not used SID (instance name) instead of databse name (without digit in the end) and vice versa.
e. You use correct Oracle installation source – 64 bit for 64 bit system and 32 bit for 32 bit system.
f. You do not have ~oracle or /etc/oracle or /etc/init.d/oracle from previous installations;
6. After installing CRS:
a. You can see crs logs (/var/log/crs/*);
b. You can see crs, css and evm in CPU after system started, and you run CRS verification script;
c. You can run crs_stat –t;
d. You run root.sh on ALL nodes, and checked CRS logs on all nodes;
e. You did not forget installation steps on some of the nodes.
f. You configured virtual IP, and configured interfaces as public and PRIVATE;
g. You can ping all VIP addresses, for example: ping rac1-vip.
h. ifconfig shows VIP addresses on correct (public) interfaces.
7. When installing DB server:
a. You installed patch for x86_64 platform after installation INSTEAD of running root.sh; then you must run root.sh . Not required on i386 platform;
b. You did not created test database here;
c. You run netca and configured listener;
d. Listener listens on correct interface (on 0.0.0.0 or on PUBLIC IP) abd VIP IP;
e. You can run crs_stat –t;
f. You relinked ORACLE with ASYNC_IO support (or remember about configuring IO slaves);
8. When you created database:
a. You can see database in crs_stat –t;
b. You can start and stop EM CONSOLE and can see database on the WEB;
c. You find ASM configuration on DB console (if you see JAVA error, try go there thru instance -> administration -> disk group. 64 it version of Oracle had a bug.)
d. You can reboot server and everything comes back;
e. You can start and stop oracle (using /etc/init.crs or rcoracle);
9. You did not set Y for cluster databases in /etc/oratab.