Collecting Diagnostic Data for OCFS2 Issues

Applies to:

Linux OS – Version: 1.4
Linux x86
Linux x86-64

Purpose

This support script is to collect diagnostic information to assist in service request troubleshooting and resolution. It gathers OCFS2 relevant information on your system and generate output in the command line output.

It only collects diagnostic information and will not do any change on your system.

Software Requirements/Prerequisites

This is a bash script and is only supported to run on RHEL/OEL 4, RHEL/OEL 5, and Oracle VM.

Configuring the Script

Ensure the execute bit on the script is turned on.

Running the Script

Run the script with root user.

It will finish in a couple of seconds with an output file named “ocfs2info.log”.

Caution

#!/bin/sh
#
# $Id: dc_ocfs2.sh,v 1.11 2009/11/26 10:05:26 rwen Exp $
#
# collect ocfs2 relevant info on localhost
#

run_cmd()
{
    local basecmd="${1%% *}"

    if type $basecmd >/dev/null 2>&1
    then
        echo "# $1"
        eval "$1"
    else
        echo "# command not exist: $basecmd"
    fi
}

dump_proc()
{
    local ENTRY="$1"
    local LEVEL="$2"

    LEVEL=$((LEVEL + 1))

    if [ $LEVEL -gt 5 ]
    then
        echo "potential loop detected."
        return 2
    fi

    if [ ! -e "$ENTRY" ]
    then
        return 1
    elif [ -f "$ENTRY" ]
    then
        run_cmd "cat $ENTRY"
    fi

    find "$ENTRY" -maxdepth 1 | while read FILE
    do
        if [ "$FILE" = "$ENTRY" ]
        then
            continue
        fi

        if [ -L "$FILE" ]
        then
            realfile=`readlink "$FILE"`
            if echo "$realfile" | grep -q '^..*'
            then
                FILE="${FILE%/*}/${realfile}"
            fi

            dump_proc "$FILE" "$LEVEL"
        fi

        if [ -f "$FILE" ]
        then
            run_cmd "cat $FILE"
        elif [ -d "$FILE" ]
        then
            dump_proc "$FILE" "$LEVEL"
        else
            echo "Ignore"
        fi
    done
}

basic_info()
{
    run_cmd "date"
    run_cmd "uname -a"
    run_cmd "id"
    run_cmd "hostname"
    run_cmd "getenforce"
}

network_info()
{
    run_cmd "ifconfig -a"
    run_cmd "route -n"
    run_cmd "grep -vE '(^;.*|^$)' /etc/resolv.conf"
    run_cmd "arp -an"
    run_cmd "netstat -s"
}

check_ocfs2()
{
    echo "Collecting package info ..." >&2

    # packages level
    run_cmd "rpm -qa --qf='%{NAME}-%{VERSION}-%{RELEASE}.%{ARCH}\n' | grep -E '(kernel|ocfs2)'"

    for i in `rpm -qa | grep ocfs2`
    do
        run_cmd "rpm -V $i"
    done

    echo "Collecting static configuration and run-time info ..." >&2

    # driver module
    run_cmd "modinfo ocfs2"
    run_cmd "lsmod | grep ocfs"

    # static configuration
    run_cmd "grep -vE '(^#.*|^$)' /etc/sysconfig/o2cb"
    run_cmd "grep -vE '(^#.*|^$)' /etc/ocfs2/cluster.conf"
    run_cmd "grep -vE '(^#.*|^$)' /etc/fstab"
    run_cmd "grep -vE '(^#.*|^$)' /etc/hosts"

    # runtime status
    run_cmd "/etc/init.d/o2cb status"
    run_cmd "/etc/init.d/ocfs2 status"

    echo "# Checking heartbeat..."
    dump_proc /sys/kernel/config/cluster/ocfs2/
    dump_proc /sys/kernel/debug/o2hb

    # disk info
    run_cmd "blkid | grep ocfs2"
    run_cmd "mount | grep ocfs2"
    run_cmd "mounted.ocfs2 -f"
    run_cmd "mounted.ocfs2 -d"

    # logs
    run_cmd "dmesg | grep -E '(Linux version|ocfs2|o2net|dlm)'"
    run_cmd "tail -n10000 /var/log/messages | grep -E '(Linux version|ocfs2|o2net|dlm)'"

    for dev in `get_dev_list`
    do
        echo "Collecting info on $dev ..." >&2
        check_one_device $dev
    done
}

get_dev_list()
{
    (blkid | grep ocfs2 | cut -d: -f1;mount | grep ocfs2 | grep -v ocfs2_dlm | awk '{ print $1 }') | sort | uniq
}

check_one_device()
{
    local DEBUGFS="debugfs.ocfs2"
    local DEV="$1"

    run_cmd "tunefs.ocfs2 -q -Q 'Block size:%B\nCluster size:%T\nNum of Nodes:%N\nVolume Label:%V\nVolume UUID:%U\n' $DEV"
    run_cmd "echo stats | ${DEBUGFS} $DEV" 2>&1
    run_cmd "echo hb | ${DEBUGFS} $DEV" 2>&1
    run_cmd "sleep 2"
    run_cmd "echo hb | ${DEBUGFS} $DEV" 2>&1

    run_cmd "echo 'stat //global_bitmap' | ${DEBUGFS} -n ${DEV}"

    run_cmd "echo 'stat //global_inode_alloc' | ${DEBUGFS} -n ${DEV}"

    run_cmd "echo 'stat //slot_map' | ${DEBUGFS} -n ${DEV}"

    run_cmd "echo 'stat //heartbeat' | ${DEBUGFS} -n ${DEV}"

    local SLOTS=`tunefs.ocfs2 -q -Q "%N\n" $DEV`
    for i in `seq --format="%04g" 0 $[${SLOTS}-1]`
    do
        INODE=//inode_alloc:$i
        EXTNT=//extent_alloc:$i
        LOCAL=//local_alloc:$i
        TRUNC=//truncate_log:$i
        JOURN=//journal:$i
        ORPHN=//orphan_dir:$i

        run_cmd "echo stat ${INODE} | ${DEBUGFS} -n ${DEV}"
        run_cmd "echo stat ${EXTNT} | ${DEBUGFS} -n ${DEV}"
        run_cmd "echo stat ${LOCAL} | ${DEBUGFS} -n ${DEV}"
        run_cmd "echo stat ${TRUNC} | ${DEBUGFS} -n ${DEV}"
        run_cmd "echo stat ${JOURN} | ${DEBUGFS} -n ${DEV}"
        run_cmd "echo ls -l ${ORPHN} | ${DEBUGFS} -n ${DEV}"
    done
}

#
# main routine starts here
#

basic_info

network_info

check_ocfs2

run_cmd "date"

Script Output

The script will generate the following output, and all the diagnostic logs will be written to ocfs2info.log.

You can check it by a text viewer or editor. At the same time, please attach it to the corresponding service request for further review.

© 2010, www.oracledatabase12g.com. 版权所有.文章允许转载,但必须以链接方式注明源地址,否则追究法律责任.

相关文章 | Related posts:

1. Collecting Diagnostic information for Oracle block corruption
2. Data Gathering for Troubleshooting RAC Issues
3. Script to Collect RAC Diagnostic Information (racdiag.sql)
4. Data gathering for troubleshooting Oracle Real Application Cluster issues
5. Data Gathering for Troubleshooting CRS Issues
6. Script to Collect Data Guard Diagnostic Information
7. Extracting Data from a Corrupt Table using DBMS_REPAIR or Event 10231
8. RMAN 11G : Data Recovery Advisor – RMAN command line example
9. How to resize an OCFS2 filesystem
10. SCRIPT TO CHECK FOR FOREIGN KEY LOCKING ISSUES