Shulou(Shulou.com)06/01 Report--

Some time ago, I mentioned some instructions on the reconfiguration of Oracle11g RAC nodes, which is more convenient and easier to manage than Oracle10g. So what does Oracle10 RAC need to do, or what do you need to pay attention to?

The first environment: add a new node, which is not described too much here, but can be added according to the official documentation.

The second environment: operating system reinstallation, here you can use the method described in the first environment, first delete the node, and then add the node. We recommend the use of official documents, but there are also some problems, such as the following:

DBCA add instance fails with:

Failed to create directory "DESCRIPTION= (ADDRESS_LIST= (ADDRESS= (PROTOCAL=tcp) (HOST=).

I'm getting the following error from DBCA when I try to add the instance on the new RAC node:

Failed to create directory "" (DESCRIPTION= (ADDRESS_LIST= (ADDRESS= (PROTOCOL=tcp) (HOST=hdoradb06.solar.cat.com) (PORT=1521) (CONNECT_DATA= (SERVICE_NAME=t3icsb_XPT) (INSTANCE_NAME=t3icsb) (SERVER=dedicated) ", LGWR ASYNC NOAFFIRM delay=0 OPTIONAL max_failure=0 max_connections=1 reopen=300 db_unique_name=" t3icsb "register net_timeout=180 valid_for= (online_logfile,primary_role)" on "hdracprd04", "Badly placed () 's: failed"

This mainly means that the environment with Dataguard is deployed. When adding nodes, dbca failed to add instances.

Reason:

DBCA does not handle remote destinations such as dataguard archive log destination. -- dbca does not handle remote destinations, such as DG log directories

Solution:

After upgrading to Oracle11.1-if you cannot upgrade temporarily, you can use tar backup and restore

Or

Stop DG, modify initialization parameters, and restart the database

Add standby log later, and configure related parameters

This is even more troublesome. Is there any other way? use tar to back up the normal node cluster and the database installation directory, restore to the reinstalled system node, change the relevant directory and file name, and execute root.sh again. Of course, the process will be a little more complicated.

Wait, we don't have to delete the instance, just like Oracle11g RAC. In fact, it is OK. Although it is not as convenient as Oracle11g, it can at least be implemented. When cleaning up cluster resources, it retains the resources of listening, ASM and database instances. For 11g RAC, please refer to the previous article explaining the matters needing attention for RAC11g node reset after OS reinstallation.

Mainly want to talk about another environment, if only the Oracle10g RAC installation directory disk is damaged, then what is more convenient to use tar backup cluster and software installation directory. This approach also effectively avoids the problems described in the second environment.

Since only the installation directory is missing, we do not need to execute root.sh script after backup and recovery through tar. After backup and recovery, check the Oracle product catalog, modify relevant directories, such as CRS_HOME/log/ "hostname", clean up the original node log information, and modify listening configuration files and TNSNAMES files.

All right, here's a reference command:

Suggestion: check the system parameters, disk attributes, etc., and check whether the undo tablespace and redo of the original node 2 exist.

[root@myrac1 oracle] # tar cvfp product2017.tar product/

Note: use root users (some groups in the cluster files are root users, please keep the permissions of the original directories and files, that is, the parameter'p')

Copy the backup file to the destination and extract it using the root user

[root@myrac2 oracle] # tar-xvf product2017.tar

After the decompression is completed, check the relevant permissions:

[oracle@myrac2 oracle] $cd product

[oracle@myrac2 product] $ls

10.2.0

[oracle@myrac2 product] $ls-lrt

Total 4

Drwxrwx--- 6 oracle oinstall 4096 Jun 20 21:39 10.2.0

[oracle@myrac2 product] $cd 10.2.0 /

[oracle@myrac2 10.2.0] $ls-lrt

Total 16

Drwxr-xr-x 44 root oinstall 4096 Jun 20 21:06 crs

Drwxr-x--- 4 oracle oinstall 4096 Jun 20 21:45 admin

Drwxrwx--- 7 oracle oinstall 4096 Nov 10 22:40 oraInventory

Drwxr-x--- 61 oracle oinstall 4096 Nov 10 23:15 db_1

Permissions are normal. Modify listening files, clean up related log directories, modify directory names, modify password and initialization parameter file names under dbs directory, check database parameters, such as Local_listener,remote_listener and other related configurations, and check again to confirm.

After the modification is complete, use the root user to execute crsctl start crs

Check the cluster running status, check the database running status, test the connection

This is for the UNIX environment, Windows, let's talk about it.

All operations or programs need to be evaluated and tested according to the existing environment, and choose a relatively best one. It is recommended to test and back up before doing the operation.

If you have a chance, you can try the second environment, using the tar method.

You are welcome to add that it is OK to slap bricks

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.