Databases Reference
In-Depth Information
Edit these files and set to the appropriate values for your environment; we recommend
referencing a healthy compute node and making the necessary adjustments. When
finished, reboot your system and again validate the configuration.
10.
The next phase of the process involves preparing the node to be added to your cluster. You
must manually perform the following tasks:
a.
Replicate Linux groups and user accounts from a healthy compute node
b.
Create directories under the /u01 mount point to store your Oracle 11gR2 RDBMS
software and Oracle 11gR2 Grid Infrastructure software and configure permissions
Setup SSH between hosts and distribute SSH keys for
dcli
to function
c.
After these tasks are complete, run
imageinfo -ver
to validate your image and compare
against another compute node in your machine:
11.
[root@cm01dbm01 ~]# imageinfo -ver
11.2.2.4.2.111221
[root@cm01dbm01 ~]#
12.
If the versions are different, you will need to patch your machine to the same version using
methods presented in Recipe 11-3.
After you have patched your node to the same image version, run
cluvfy
with the
following syntax to ensure your server is ready for an Oracle cluster installation. The
command will compare your current, reimaged server with a healthy server in your
cluster:
13.
[root@cm01dbm01 ~]# cluvfy stage -post hwos -n cm01dbm01,cm01dbm02 -verbose
14.
Next, verify the peer compatibility between your nodes. In the following example,
cm01dbm01
is the reimaged server and
cm01dbm02
is a healthy server:
[root@cm01dbm01 ~]# cluvfy comp peer -refnode cm01dbm02 -n cm01dbm01 -orainv \
> oinstall -osdba dba | grep -B 3 -A 2 mismatched
15.
After resolving issues, you can begin adding your reimaged server back into the cluster
using
addNode.sh
from a functional node:
[root@cm01dbm02 bin]# ./addNode.sh -silent "CLUSTER_NEW_NODES=cm01dbm01" \
> "CLUSTER_NEW_VIRTUAL_HOSTNAMES=cm01dbm01-vip"
Follow a similar process to clone your database home from a healthy server to the compute server, as outlined in
the
Exadata Owner's Manual
.
How It Works
At a high level, reimaging a compute node server entails booting from an external USB drive that contains a factory
boot image and using Exadata's boot process to configure your network information and install an Exadata image
onto your compute node. After a successful compute node reinstall, several manual steps must be performed by the
Exadata DMA to bring the compute node to an operational state.
Some of the post-reimage steps are specific to Oracle Exadata, whereas others could be more accurately classified
as being related to adding a database server than to an Oracle RAC cluster. These steps are documented very well in
Search WWH ::
Custom Search