Recovery for data
loss, host failure, or media failure (critical): These procedures
enable you to recover from critical failures that involve actual data
loss. In all cases, these procedures involve making sure your state is
consistent across all installations.
Depending on the type
of loss, they can involve recovering any combination of the following
types of files:
- Oracle software files and configuration files
- Database files
- Oracle Calendar database files
- Oracle system files
Recovery
for process crashes or system outages (non-critical): These procedures
involve restarting processes that have stopped or failed. These procedures
generally do not involve restoring data, but restoring data would be necessary
in cases where a process crashes during a write operation to a configuration
file and ended up corrupting that file.
In the case of a host or disk failure, you may need to restore
to another computer. If you have a complete host backup, then you can use it
to restore your entire host to a previous configuration. If you do not have
a complete host backup, perform the following steps:
1.
Prepare
a new host that has identical system configuration as the original host.
Use the configuration record to ensure that there is an identical system
configuration.
2.
Make
sure that the IP address, hostname, and aliases for the OCS hosts in /etc/hosts
are identical to the old /etc/hosts
file. If you have lost the disk containing your /etc/hosts
file, the you may need to perform this step when restoring to the same
host.
3.
Check port usage on the new host. Make sure there
are not any processes using the same ports as the Oracle Collaboration
Suite installations you are about to restore. If there are, you must reconfigure
these processes to use different ports before you begin restoring your
Oracle Collaboration Suite installations.
4.
Create an operating system user that is identical
to the user who installed OCS on the original host. The following
attributes should be the same:
- User name
- User ID
- Group name
- Group ID
- Environment profile
- Shell
The user may have the same password or a different password
than the original user.
5.
Set file permissions. Begin by running the following
command as root:
$ORACLE_HOME/root.sh
6.
Continue the recovery process for the specific type
of OCS tier. Your new system now looks exactly like the failed system.
If you are restoring your Infrastructure to the same host
after a disk failure, then you may not need to restore all of the Infrastructure.
For example, if the failed disk only contains the Metadata Repository, then
just recovery that database. To restore as quickly as possible, only restore
the portions of the Infrastructure that have been lost due to the disk failure.
To restore your Infrastructure, perform the following steps:
1.
If the Infrastructure is still running, shut it down.
Click here
to reference "Shutting Down Oracle Collaboration Suite" OBE
for complete details.
2.
If you have lost any disks with system configuration
files, ensure that the system configuration is restored. For example,
if the disk containing the /etc/hosts
file has failed, ensure that the IP address, hostname, and aliases for
the OCS hosts in /etc/hosts
are identical to the old /etc/hosts
file.
3.
If you have lost a disk containing the Oracle system
files, then restore these files. These files can be on two drives, with
the /var/opt/oracle or
/etc directory on one disk
and the oraInventory directory
on another.
4.
If
you have lost a disk containing the Infrastructure home directory, then
restore this directory using the following steps:
- Make sure that the filesystem where you will create the directory has
enough space to hold the Infrastructure.
- Create an empty directory using the same mount point and full path as
the original Infrastructure home directory. Do not use symbolic links
anywhere in the path.
- Make sure the directory is owned by the same user and group as on the
original host.
- Restore the Infrastructure home directory from your most recent complete
backup. Be sure your method of restoring the files preserves the original
owner, group, permissions, and timestamps.
- Restore the configuration file backup from your most recent partial
online backup.
.
5.
If you have lost a disk containing the Metadata Repository,
then restore and recover the Metadata Repository, using the normal database
recovery techniques. If it is not possible to do a full recovery, because
of the absence of online redo logs, then perform a point-in-time recovery.
6.
If
you have lost a disk containing the OID, then restore and recover the
OID, using the normal database recovery techniques. If it is not possible
to do a full recovery, because of the absence of online redo logs, then
perform a point-in-time recovery.
7.
Start
your Infrastructure. Click here
to reference "Starting Up Oracle Collaboration Suite" OBE for
complete details.
As with an Infrastructure recovery, recover only the files
that have been lost due to a disk failure. To restore an Information Store,
perform the following steps:
1.
Before restoring the Information Store, initialize the
following tiers:
-
Shut down
any middle-tier applications that use the Information Store that you are
restoring. You can leave other applications running.
- If necessary, shut down the Information Store that you are restoring.
- If necessary,
start your Infrastructure.
2.
If
you have lost any disks with system configuration files, ensure that the
system configuration is restored.
3.
If you have lost a disk containing the Oracle system
files, then restore these files.
4.
If you have lost a disk containing the Information Store
home directory, then restore this directory using the same steps used
when recovering the Infrastructure.
5.
If you have lost a disk containing the Information Store
database, then restore and recover the database, using the normal database
recovery techniques. If it is not possible to do a full recovery, because
of the absence of online redo logs, then perform a point-in-time recovery.
6.
If you have lost a disk containing Oracle Files BFILEs,
use your preferred operating system command to restore the BFILEs.
7.
Restart these components of Oracle Collaboration Suite,
as the other OCS processes should already be started:
- Start the listener and instance of the recovered Information
Store
- Start any middle tier applications that use the Information Store
As with
an Infrastructure recovery, recover only the files that have been lost due to
a disk failure.
To restore your middle tier, perform the following steps:
1.
Before restoring the Middle Tier, initialize the following
tiers:
-
If there are any application running on the host to be recovered, shut
then down.
- If necessary, start your Infrastructure and Information Stores.
2.
If
you have lost any disks with system configuration files, ensure that the
system configuration is restored.
3.
If you have lost a disk containing the Oracle system
files, then restore these files.
4.
If you have lost a disk containing the middle tier home
directory, then restore this directory using the same steps used when
recovering the Infrastructure.
5.
If the failed disk includes the Calendar database or
configuration files, use the unidbrestore
utility to restore the Calendar database.
The
unidbrestore utility restores
calendar server node and configuration information from a backup created by
unidbbackup. Archived backups
should be managed to ensure full data recovery capabilities without sacrificing
large amounts of disk space. Remove backups that are no longer needed.
unidbrestore
restores only the database and configuration files. Calendar data stored in
OID must be restored separately. If you have any reason to expect that inconsistencies
may exist between the data in the calendar server and that in the directory
server, use the unidsdiff and
unidssync utilities to identify
and resolve all discrepancies. unidbbackup
is the complementary utility to unidbrestore.
To
restore the calendar database, perform the following steps:
1.
Use
unidbrestore to restore
the database and configuration files. For example, to restore node 45
from the /backups/cserver/jan.7.04
backup to the directory $ORACLE_HOME/ocal,
issue the following command:
Bookmark
