Database Reference
In-Depth Information
fi
read thisSID?"Enter the name of the database to be monitored again: "
else
thisSID=`print $@ | awk '{print$NF}' | tr '[A-Z]' '[a-z]'`
fi
BO_NAME=scripted_blackout_${thisSID}
CleanUpFiles
EndBlackout
ExitCleanly
Section 2: EM CLI and Veritas Cluster Server
Veritas Cluster Servers (VCS)
1
are robust fail-over mechanisms deployed as an alternative to RAC One-Node. With VCS,
critical resources like database instances are identified and monitored. If a failure is noticed, VCS will unmount drives
containing the binaries and database files, remount them on the other node of a cluster, and restart the services.
OEM agents must be installed on a static drive on each host within the cluster. They never fail-over with the VCS
resource group because they are strictly associated with a single host. This poses a problem with VCS fail overs. How
do you keep track of the targets when the database, its Oracle home, and all of its files move to another host?
■
many consider it a best practice on VCS to install the agents on the virtual hostname. refer to “how to Install 12c
agent on Virtual hostname / Virtual Ip address?” (Doc ID 1469995.1) for details.
Tip
The EM CLI
relocate_targets
verb transfers responsibility for target tracking to another agent when the targets
are relocated by the VCS. This section describes how that process works before showing you the shell script itself.
Program Logic
This script gathers data from the environment and decides which targets are moving and which agents are affected.
This script's scalability and flexibility come from queries of three systems: the VCS cluster, the OEM repository, and the
EM agent. Note that syntax is slightly different for relocating responsibility between agents for database and listener
targets. We'll generate a separate set of workfiles for generating the EM CLI argfile that actually performs the work.
Veritas Cluster Server
VCS clusters are organized into groups that drill down to the individual database and listener targets. The hosts in the
cluster are at the highest level. VCS heartbeats monitor one or more critical resources on the active node and when
necessary relocate all the members
2
of a group from the current ONLINE node to the other node that is currently
OFFLINE. We'll use this knowledge to determine the host and agent names we need to manage.
2
Members on a database server typically include databases and listeners using the same Oracle home. The drives related to these
targets are all members of the group. Filesystems containing the Oracle base, the archive destination, the oradata filesystems, and
other related data are all included in the fail-over group.
1
Search WWH ::
Custom Search