4 High Availability

This chapter describes issues related to highly available topologies. This chapter contains the following issues:

Section 4.1, "OracleAS Disaster Recovery: Discover Topology Command"
Section 4.2, "OracleAS Disaster Recovery: Real Application Clusters Database Supported"
Section 4.3, "OracleAS Backup Recovery: prefix too long Errors During an Image Backup"

4.1 OracleAS Disaster Recovery: Discover Topology Command

If you run the discover topology command on a node that contains more than one Oracle home, and one of the Oracle homes is invalid for some reason (that is, the Oracle home does not appear in the Oracle Universal Installer), the discover topology command generates a warning:

ASGCTL> discover topology oidpass=welcome1
Discovering topology on host "hasun1" with IP address "123.45.67.89"

hasun1:7890
 Connecting to the OID server on host "hasun12vip1.mydomain.com"
   using SSL port "636" and username "orcladmin"
 Getting the list of databases from OID
 Gathering database information for SID "orcl" from host
                                               "hasun12vip1.mydomain.com"
 Getting the list of instances from OID
 Gathering instance information for "immr.hasun12vip1.mydomain.com" from host
                                               "hasun12vip1.mydomain.com"
 Gathering instance information for "asmid.haqadr01.mydomain.com" from host
                                               "haqadr01.mydomain.com"
      **********  WARNING  **********
hasun1: -->ASG_IAS-15779: Error getting instance information for instance
 "asmid.haqadr01.mydomain.com" from host "haqadr01.mydomain.com".  This instance
 will be excluded from the topology.xml file
drmt: -->ASG_IAS-15632: The home that contains instance 
 "asmid.haqadr01.mydomain.com" could not be found
drmt: -->ASG_DUF-4950: An error occurred on host "drmt" with IP "130.35.45.23" and
 port "7890"
      ********  END WARNING  ********
The topology has been discovered. A topology.xml file has been written to each
home in the topology.

To work around this issue, delete the entry for the invalid Oracle home from the Inventory.xml file in the oraInventory directory, then rerun the discover topology command.

4.2 OracleAS Disaster Recovery: Real Application Clusters Database Supported

Table 1-3 in the Oracle Application Server High Availability Guide incorrectly stated that OracleAS Disaster Recovery does not support OracleAS Infrastructure in active-active topologies. OracleAS Disaster Recovery does support OracleAS Infrastructure in active-active topologies, as well as active-passive topologies. In OracleAS Infrastructure active-active topologies, the OracleAS Metadata Repository runs on a Real Application Clusters database.

The following table shows the updated Table 1-3 (bold text shows the updates):

Table 4-1 Service level requirements and architecture choices

Business Requirements			Architecture Choices
Local High Availability	Scalability	Disaster Recovery	Instance Redundancy	Disaster Recovery
N	N	N	Base	N
Y	N	N	Active-passive	N
N	Y	N	Active-active	N
N	N	Y	Base	Y
Y	Y	N	Active-active	N
Y	N	Y	Active-passive	Y
N	Y	Y	Active-active (middle tier) Base (Infrastructure)	Y
Y	Y	Y	Active-active (middle tier) Active-passive and active-active (Infrastructure)	Y

Note:

OracleAS Disaster Recovery supports the base, active-passive, and active-active Infrastructure architectures. For additional scalability in a base, active-passive, or active-active architecture, extra computing power can be added to the infrastructure hardware (for example, high capacity CPUs, more memory)

4.3 OracleAS Backup Recovery: prefix too long Errors During an Image Backup

The following command and its options are used to create an Image Backup during the set of Oracle Application Server Backup and Recovery operations:

bkp_restore.sh -m node_backup -o image_backup -P <archive path>

This command uses the first tar executable found in the user's $PATH. This command may fail with prefix too long errors due to long pathname limitations in the /usr/bin/tar executable that is installed by default.

This issue is documented in Oracle Bug #4599526 (LOHA :PREFIX TOO LONG WITH NODE_BACKUP (LOHA)).

The workaround is to install a third-party tar, the GNU tar, into /usr/local/bin and place /usr/local/bin within the $PATH of the user, before /usr/bin and before the Image Backup command is executed. Oracle has verified that the GNU tar does not have any long pathname limitations. Any other available tar that does not have long pathname limitations can also be used.