 |
Index for
Section 8 |
|
 |
Alphabetical
listing for C |
|
 |
Bottom of
page |
|
clu_check_config(8)
NAME
clu_check_config - Check cluster subsystems on the local member
SYNOPSIS
/usr/sbin/clu_check_config [-blqv] [-s subsystem[,subsystem]...]
/usr/sbin/clu_check_config -s
FLAGS
-b Boot mode. The clu_check_config command performs a quick check of
the specified subsystems or applications. (For simplicity, the
rest of this reference page uses the term subsystem to refer both
to subsystems and applications.)
Each diagnostic utility is called in turn and given a fixed
amount of time to return status. The clu_check_config command
monitors the execution of each utility and terminates the utility
if it does not return within the allotted period of time.
Terminating a diagnostic utility automatically sets the
clu_check_config exit status to -1.
Standard out, stdout, and standard error, stderr, are logged to
/cluster/admin/clu_check_log_full_hostname_of_member.
By default, each system runs clu_check_config -b the first time
it is booted as a cluster member. You can also run this command
in multiuser mode.
-l Local logging. Standard out, stdout, and standard error, stderr,
are logged to ./clu_check_config.log.
If -b and -l are both specified, clu_check_config performs a
boot-mode check, but writes the log file to
./clu_check_config.log.
-q Quiet mode. Redirect both stdout and stderr to /dev/null.
However, if -q is specified with -b or -l, redirect stdout and
stderr to a log file.
-s subsystem[,subsystem]...
Subsystems. A comma-separated list, with no spaces, of subsystems
to check. For each specified subsystem, clu_check_config calls
the appropriate diagnostic utility located in the
/cluster/admin/clu_check_config directory.
If you run clu_check_config without specifying the -s flag, all
diagnostic utilities are called. For example, clu_check_config -l
or clu_check_config -vb call all diagnostic utilities.
-s Subsystem list. Display a list of valid subsystem names for use
with the -s flag.
-v Verbose. Send additional information to stdout. Note that if you
specify -bv or -lv, you get verbose output, but it is redirected
to a log file, not sent to stdout.
DESCRIPTION
The clu_check_config command checks the configuration of the cluster, from
the point of view of the local member, by calling one or more diagnostic
utilities located in the /cluster/admin/clu_check_config directory. (A
diagnostic utility can be associated with any cluster subsystem, UNIX
subsystem, or application.)
You must be root to run clu_check_config.
When a new member is added to the cluster, clu_check_config -b is
automatically run at the first boot of that member. To check the health of
the entire cluster, run this command on each cluster member at any time.
The clu_check_config command provides an execution environment for these
diagnostic utilities. It does this by providing the following:
· Diagnostic utility scheduling and monitoring - The clu_check_config
command invokes each diagnostic utility as a background process,
redirecting stdout and stderr as indicated by the optional arguments
provided to clu_check_config.
· Logging and log file maintenance - Each invocation of clu_check_config
that specifies logging (-b or -l) appends a log to the appropriate log
file. When invoked without any options, clu_check_config writes to
stdout.
Date and time stamps separate logs within a log file. Log files
contain messages from clu_check_config and any diagnostic utilities
called by clu_check_config.
· Command-line option processing - Controls which subsystems are
checked, what is displayed to the screen, and what is written to log
files.
The remainder of this section contains guidelines for creating diagnostic
utilities.
Creating Diagnostic Utilities
The primary purpose of each diagnostic utility is to provide the current
status of a subsystem or application. Each time it is invoked, a diagnostic
utility must exit with a valid return status:
0 Good Configuration - The configuration of the subsystem or
application is valid. Diagnostic utilities run on a properly
installed and configured cluster should exit with a status of 0.
1 Bad Configuration - The configuration of the subsystem is not
valid.
-1 (255) Warning/Other - Either the validity of the configuration cannot
be determined or the configuration is valid but not optimal.
A diagnostic utility should return a Warning/Other exit status
when a required resource is not available. For example, the
utility cannot find a file, configuration value, process, or even
another subsystem or application that is required to perform the
diagnosis of the specified subsystem or application. The utility
can also return this status when it determines that the subsystem
configuration, while valid, might not be optimal, highly
available, or logical.
All diagnostic utilities called by clu_check_config must be placed in the
/cluster/admin/clu_check_config directory. Each utility must be a file or
a symbolic link to a file that is executable by root. Diagnostic utilities
should have owner and group set to bin and a mode of 700.
In order to facilitate a logical ordering of diagnostic utilities, a
diagnostic utility file naming convention is enforced. This naming
convention is similar to that used in the /sbin/rc#.d directories.
· The first two characters of the file name are a sequence number. If
you are not sure what number to use, use the sequence number from
/sbin/rc3.d for the subsystem or application diagnostic utility.
· The remainder of the file name is arbitrary. If the subsystem has an
analog in /sbin/init.d, you can use a similar name here. For example:
/cluster/admin/clu_check_config/50check_cdsl_config
Do not display the sequence number portion of the file name as part of
any informational or error message.
When invoked without a list of subsystems, clu_check_config calls each
diagnostic utility in the order that an ls command would list the contents
of the /cluster/admin/clu_check_config directory (in ascending ASCII sort
order).
Each diagnostic utility must provide certain features. A utility must:
· Accept all command-line flags passed to it by clu_check_config.
· Process the -b command-line flags and limit its execution as defined
by the boot-mode definition described later.
· Save the current state of any subsystem the utility needs to modify in
order to determine the configuration of the subsystem the utility is
diagnosing.
· Print information messages to stdout.
· Print all errors and warnings to stderr.
· Restore any modifications made to any subsystems during the utility's
execution.
· Remove persistent temporary data and terminate child processes.
· Exit with a valid exit status.
· When passed an unknown command-line flag, execute in its default mode
(as though it were passed no flags).
Each diagnostic utility has the option of providing additional
informational messages when passed the verbose, -v, flag.
Diagnostic Utility Mode Definitions
Each diagnostic utility can have one or more modes of operation. The
current supported modes are boot mode and default mode:
Boot Mode Boot mode is indicated by the presence of the -b command-line
flag. Boot mode's purpose is to provide a quick (within 20
seconds) determination of whether a subsystem's status is good,
bad, or indeterminate. This limited amount of time might mean
that a complete analysis of a subsystem is not possible. The
diagnostic utility should do the best job it can do within this
time restriction.
Default Mode
Default mode is indicated by the absence of any command-line
arguments, other then the verbose, -v, flag. A utility's default
mode should provide a thorough analysis of the subsystem
indicating whether the configuration is good, bad, or
indeterminate. In addition, the level of detail provided by
default mode should be greater then that provided in boot mode.
Diagnostic Utility Output Recommendations
Each diagnostic utility displays output as follows:
· At startup, print an information message to stdout stating what
subsystem or application the utility is going to diagnose.
· While executing, print any errors, warnings, or abnormal findings to
stderr.
· While executing, if passed a -v option by clu_check_config, the
utility has the option of printing additional informational messages
to stdout.
· At completion, print an information message stating the results of the
diagnosis to stdout.
EXAMPLES
The following examples show how combinations of flags determine the output
and logging of the clu_check_config command.
To check all subsystems and applications known to clu_check_config (those
that have diagnostic utilities in the /cluster/admin/clu_check_config
directory) and display verbose informational messages or errors to stdout
and stderr:
# clu_check_config -v
To check all subsystems and applications known to clu_check_config, display
no informational messages or errors, and log verbose output to
./clu_check_config.log:
# clu_check_config -qvl
To obtain only the exit status of clu_check_config:
# clu_check_config -q
To perform a quick check of all subsystems and applications known to
clu_check_config and log verbose output to
/cluster/admin/clu_check_log_full_hostname_of_member:
# clu_check_config -bv
EXIT VALUES
The exit values for clu_check_config are:
0 Good Configuration - All diagnostic utilities returned success.
The configuration of the tested subsystems or applications is
valid.
The clu_check_config command should exit with 0 on a cluster that
has been installed and configured as recommended.
1 Bad Configuration - One or more diagnostic utilities did not
return 0. The configuration of one or more subsystems or
applications is not valid.
-1 (255) Warning/Other - One or more diagnostic utilities returned an exit
status other than 0 or 1. Either the validity of the
configuration could not be determined or the configuration is
valid but not optimal.
FILES
/usr/sbin/clu_check_config
The command path.
/cluster/admin/clu_check_log_full_hostname_of_member
Log file when -b is specified.
./clu_check_config.log
Log file when -l is specified.
/cluster/admin/clu_check_config
Directory containing the diagnostic utilities.
RELATED INFORMATION
TruCluster Server Cluster Installation
TruCluster Server Cluster Administration
|
Index for
Section 8 |
|
|
Alphabetical
listing for C |
|
 |
Top of
page |
|