Top
PRIMECLUSTER Wizard for Oracle 4.2 Configuration and Administration Guide

4.5 clgetoralog - Collect Troubleshooting Information

User

Super user

Format

/opt/FJSVclora/bin/clgetoralog [ -n RMSconfiguration ] [ -d output_directory ] [ -a ]

Function

clgetoralog collects Oracle information for troubleshooting. The “clgetoralog” command collects Oracle setup information and log files on a PRIMECLUSTER system.
The information will be collected in the “<hostname>_<yymmddHHMMSS>_clgetoralog.tar” file.


The “clgetoralog” command collects

  • initialization parameter file
    $ORACLE_HOME/dbs/*.ora
    $ORACLE_BASE/admin/$ORACLE_SID/pfile/*.ora

  • Oracle alert log
    background_dump_dest/*.log
    $ORACLE_BASE/admin/$ORACLE_SID/bdump/*.log
    $ORACLE_HOME/rdbms/log/*.log
    $ORACLE_HOME/dbs/*.log
    Under $ORACLE_BASE/diag/

  • listener.ora, tnsnames.ora
    $ORACLE_HOME/network/admin/*.ora

  • Listener log
    $ORACLE_HOME/network/log/*.log

  • Oracle Clusterware Information
    $ORA_CRS_HOME/crs/admin/*
    $ORA_CRS_HOME/crs/log/* ($ORA_CRS_HOME/log/<hostname>/crsd/*)

  • V/L, configuration, and log files of PRIMECLUSTER Wizard for Oracle

  • V/L and log files (switchlog) of RMS

  • V/L, configuration, and patch information of PRIMECLUSTER

  • System information
    uname -a (system basic information)
    uptime (system operating time)
    psrinfo -v
    /usr/platform/`arch -k`/sbin/prtdiag (*1)
    cat /etc/release (Solaris version)
    cat /var/sadm/system/admin/CLUSTER (Solaris software group)
    cat /var/sadm/ptf/newprev/Rcontents (PTF information)
    grep -h “Install Solaris” /var/sadm/install_data/*OK (RSPC information)
    pkgparam -v FJSVbse (ESF information)
    isainfo -b, isainfo -v (instruction set architectures information)
    sysdef -i (kernel parameter)
    ipcs (shared resources)
    swap -l (swap information)
    ptree -a, ps -ef (process information)
    ntpq -p (NTP information)
    /var/adm/messages files

In the environment that applied the patch 915058-02 or later of PRIMECLUSTER Wizard for Oracle, the following information is gathered.

  • System information (Solaris 10 only)

    ptree -a, ps -efZ, ps -efL -o (process information)

    zonename, zoneadm list -vc (zone information)

(*1) It is not gathered in Oracle Solaris 9 Containers.

Information

The fjsnap command can also be used to collect troubleshooting information.

Parameter
[ -n RMSconfiguration ]

Specify RMS configuration name for RMSconfiguration.
If this option is specified, clgetoralog command collects RMS configuration information in addition to the information described in “Function”. When the cluster application has been built by using GUI (userApplication Configuration Wizard), the RMS configuration name is “config”. The RMS configuration name can be confirmed by the following command while RMS is running.

# /opt/SMAW/bin/hvdisp -a | grep Configuration
Configuration: /opt/SMAW/SMAWRrms/build/config.us

If this option is not specified, clgetoralog command collects only information described in “Function“.

[ -d output_directory ]

Specify the directory where information is collected. Then, the command will store the “<hostname>_<yymmddHHMMSS>_clgetoralog.tar” file in the directory. If the directory specified does not exist, it will automatically be created.
If this option is omitted, the command will store the “<hostname>_<yymmddHHMMSS>_clgetoralog.tar” file under the default directory (/var/tmp/).

[ -a ]

The option will collect the following information in addition to the information described in [Function]:

  • Group of currently active RMS configuration files (all RMS configuration files with the -n option)

  • RMS log files (/var/opt/reliant/log/*)

  • RMS environment variable configuration file (/usr/opt/reliant/bin/hvenv.local)

  • /var/adm/messages*

  • /etc/hosts, /etc/inet/hosts

  • /etc/passwd

  • /etc/services, /etc/inet/services

  • /etc/system

  • /etc/vfstab

  • showrev -p

If this option is omitted, the command will collect information described in [Function] only.

Exit status

0: Normal termination
Non-0: Abnormal termination

Note

This command gets the value of background_dump_dest or diagnostic_dest parameter from the initialized parameter file or the server parameter file, and searches the Oracle alert log. However, if these parameter files can not be read(e.g. the server parameter file exists on a shared disk that is not mounted), the alert log is not also collected.
To prevent it, define the alert log in /opt/FJSVclora/etc/clgetoralog.conf using the full path so that you can acquire the alert log. Where the alert log is stored is defined in background_dump_dest parameter. (background_dump_dest/diagnostic_dest parameter can be found in the initialized parameter file or the server parameter file. It can be also displayed by SQL “show parameter dest”.)

The following example shows how to define /opt/FJSVclora/etc/clgetoralog.conf:

COLLECTOBJ=/oracle/admin/orclsid/bdump/alert_orclsid.log

* background_dump_dest = /oracle/admin/orclsid/bdump

Start with “COLLECTOBJ=” and specify the file name using the full path after “=”.
Wild card characters “*” cannot be used in the middle of the path name.

COLLECTOBJ=/oracle/admin/orclsid/bdump/*.trc	[OK]
COLLECTOBJ=/oracle/admin/*/bdump/*.trc	[NG]

If wild card characters are frequently used, and the data size becomes large, troubleshooting information might not be collected.