Database Reference
In-Depth Information
A script to validate RAID is also available and run weekly to ensure the RAID configurations of the box. Listing 7-29
shows a snippet of the raid-check script.
Listing 7-29.
Raid Check Script
#!/bin/bash
#
# This script reads it's configuration from /etc/sysconfig/raid-check
# Please use that file to enable/disable this script or to set the
# type of check you wish performed.
# We might be on a kernel with no raid support at all, exit if so
[ -f /proc/mdstat ] || exit 0
All the scripts listed are enabled by default on the ODA, but the output is e-mailed to the internal root account
using the mail utility. In order to look at the outputs, you need to log on to the box as the root user and execute the
mail utility or set up mail redirection as described in MOS Note 405229.1.
Reactive Checks
Proactive checks have been the focus of this chapter so far. Things can happen, however, and it is important to
understand various reactive procedures that are needed or embedded in the ODA. Diagnostics are a very complex
process, and thus not all forms of reactive checks can be covered as part of this section, but we do cover some of the
more useful ones.
Diagnostic Collection
The ODA provides an easy way to collect and manage various diagnostic information in the form of the
oakcli
manage diagcollect
command. The command allows for easy access to all log files and diagnostic information
across various ODA, Clusterware, ASM, and RDBMS log files, and on multiple nodes. Listing 7-30 shows how to get
the help text for the command.
Listing 7-30.
ODA Diagnostic Collection Utility
./oakcli manage diagcollect -h
Usage:
oakcli manage diagcollect [options]
options:
--all - For collecting all diag info (Default option),
excluding adr, chmos
--crs - For collecting crs diag information
--patch - For collecting ODA Patching logs
...
Diagnostic information can be trimmed and, depending on the type of problem, only specific information can be
collected or excluded. Diagnostic data when paired with TFA can help provide Oracle with a lot of information,
which in turn can help debug and diagnose a problem pretty quickly.
