tsysinfo

#!/bin/bash
# 
# the only dependancy on bash ^^ is in function _socket_send() - not sure how to remove that!
# 
# 
# tsysinfo  v1.0.1 ('Terse System Information')
#
#   Report back on certain very basic but vital aspects of system health
#   Each as a really short one-liner, with maximum terseness / brevity
#
# Example Output:
#
#   https://gist.github.com/dreamcat4/21b67ffe135546697b5411ceb26246f1
#
#
# Version History:
# 
# v1.0.0 (initial version)
#   Created by: Dreamcat4 (dreamcat4@gmail.com)
#
# v1.0.2 - added further inline documentation
# v1.0.3 - added link to example output
# v1.0.4 - lshw requires elevated permissions
# v1.0.5 - replace lshw with lsscsi
# v1.0.6 - 'implement socket-listen' and 'socket-send'
#          lets communication from within a container. to query hardware on the host system
# v1.0.7 - improve output of cpu-cores to be Nc/Nt
# v1.0.8 - fix zfs-health subcmd - broken parsing of 'zpool status'
# v1.0.9 - changed license to apache v2.0
#
#
# Notes:
#   v1.0.0
#   Created and tested on ubuntu 18.10 and higher
#   requires a few optional packages for certain properties
#   otherwise a status of "n/a" will be returned
#
# Contribution guidelines:
#   Released under public domain and the general Apache v2.0 license
#   If making further changes or improvements then just fork in github
#   Also can append a new entry to the version history as a contributor.
#   
#
# Dependancies:
#   (ubuntu) It's recommend to install the following packages:
#
#     apt install sysstat lm-sensors lsscsi smartmontools socat bash
#
#   and also to then run the following configuration script:
#
#     sensors-detect
#
#
#
#

_program="$(basename $0)"

_msg_not_available="n/a"

err()
{
    _rc="$1"
    shift
    echo "${_program}: error: $@"
    exit $_rc
}

warn()
{
    echo "${_program}: warning: $@"
}


info()
{
    echo "${_program}: info: $@"
}


_try_sudo()
{
    if [ "$(id -u)" = "0" ]; then
        unset _sudo
    else
        if [ "$(command -v sudo)" ]; then
            if sudo su root -c "" 2> /dev/null; then
                export _sudo="sudo"
            else
                return 1
            fi
        else
            return 1
        fi
    fi
}


_cat_help()
{
    cat <<- EOF

    $_program:
    Print out basic system specs, and system health diagnostic information.

    usage:

    $_program [subcommand|--help]

    subcommands:

    cpu-cores
    cpu-temp
    cpu-usage
    disk-health
    disk-usage <mountpoint>
    disks
    disks-worst-temp
    dmesg-health
    docker-status
    fan-spin <2,4,5>
    kernel
    memory-usage
    os
    platform
    systemd-health
    systemctl-status <service>
    uptime
    zfs-health
    socket-listen <unix.sock>
    socket-send <unix.sock> <subcommand> [args]

    --debug, -d:
    Enable shell debugging
    --help, -h:
    Display this message and exit

EOF
}










_platform()
{
    # report the platform architecture, for example 'x86_64'
    uname --hardware-platform
}

_kernel()
{
    uname --kernel-name --kernel-release
}

_uptime()
{
    uptime -p | sed -e "s/^up //g"
}

_os()
{
    # for ubuntu
    if command -v lsb_release > /dev/null; then
        lsb_release -d | cut -f2
    else
        echo "$_msg_not_available"
    fi
}

_cpu_cores()
{
    # report back the number of cores(threads)
    if [ -e "/proc/cpuinfo" ]; then
        printf "$(grep -m1 'cpu cores' /proc/cpuinfo | sed -e 's/.*: *//g')c/$(nproc --all)t\n"
    else
        echo "$_msg_not_available"
    fi
}

_cpu_usage()
{
    # report the current and average whole-cpu usage, in percent
    # requires the 'sysstat' package, for the command 'mpstat'
    if [ "$(command -v mpstat)" ]; then

        _idle_avg="$(mpstat | tail -1 | sed -e "s/  */ /g" | cut -d" " -f12)"
        _load_avg="$(echo "100 - $_idle_avg" | bc)"

        _idle_cur="$(mpstat 1 1 | tail -1 | sed -e "s/  */ /g" | cut -d" " -f12)"
        _load_cur="$(echo "100 - $_idle_cur" | bc)"

        echo "${_load_cur}% cur / ${_load_avg}% avg"

    else
        echo "$_msg_not_available"
    fi
}

_cpu_temp()
{
    # report the current cpu whole-package temperature
    # requires the package 'lm-sensors', for the command 'sensors' (you must also run 'sensors-detect' too)

    if [ "$(command -v sensors)" ]; then
        sensors | grep -i "package" | sed -e "s/.*: *+//g" -e "s/ .*//g"

    else
        echo "$_msg_not_available"
    fi
}

_fan_spin()
{
    # report on fan status, alert if a specified fan sensor is reporting 0rpm 
    # otherwise report the fan rpms, in the order given respectively

    # takes a single argument, comma separated list of fans e.g. "1,2,3"
    # requires the package 'lm-sensors', for the command 'sensors' (you must also run 'sensors-detect' too)
    _fans="$(echo "$1" | sed -e "s/,/ /g" -e "s/fan//g")"

    if [ "$(command -v sensors)" ]; then

        _sensors_fans="$(sensors | grep -i fan)"

        if [ "$_sensors_fans" ]; then

            unset _msg
            for _fan in $_fans; do
                _fan_rpm="$(echo "$_sensors_fans" | grep -i "^fan${_fan}:" | sed -e "s/  */ /g" | cut -d " " -f2)"

                if [ "$_fan_rpm" ]; then

                    if [ "$_fan_rpm" -eq 0 ]; then
                        _msg="${_msg}, fan${_fan} has failed"

                    else
                        _fan_rpms="${_fan_rpms}, ${_fan_rpm}rpm"
                    fi

                else
                    _msg="${_msg}, fan${_fan}: no data"
                fi
            done

            if [ "$_msg" ]; then

                if echo "$_msg" | grep -q -i "failed"; then
                    echo "CRITICAL: ${_msg#, }"

                else
                    echo "${_msg#, }"
                fi
            else
                echo "ok (${_fan_rpms#, })"
            fi
        else
            echo "$_msg_not_available"
        fi
    else
        echo "$_msg_not_available"
    fi
}

_memory_usage()
{
    # report back the amount of real memory used vs total ram, in a human readable format
    _free_output="$(free -h --giga | sed -e "s/  */ /g")"

    _mem_used="$(echo "$_free_output" | grep "Mem:" | cut -d" " -f3)"
    _mem_total="$(echo "$_free_output" | grep "Mem:" | cut -d" " -f2)"

    echo "${_mem_used}/${_mem_total}"
}

_dmesg_health()
{
    # Reports back the number of errors and warnings found in the dmesg log (or 'ok' if there are none)
    # requires the 'dmesg' command. its good for detecting lower level hardware errors
    _dmesg="$(dmesg)"
    _dmesg_warnings="$(echo "$_dmesg" | grep -i "] warning: ")"
    _dmesg_errors="$(echo "$_dmesg" | grep -i "] error: ")"

    if [ "$_dmesg_warnings" ] || [ "$_dmesg_errors" ]; then
        _dmesg_num_warnings="$(echo "$_dmesg_warnings | wc -l")"
        _dmesg_num_errors="$(echo "$_dmesg_errors | wc -l")"

        _output_msg="${_dmesg_num_warnings} warnings, ${_dmesg_num_errors} errors"

    else
        echo "ok"
    fi
}

_disks()
{
    # report the total number of real physical disks that are currently detected attached to the system
    # requires the package 'lsscsi' for the command 'lsscsi'

    _num_disks="$(lsscsi | wc -l)"

    if [ $_num_disks -gt 0 ]; then
        echo "$_num_disks"
    else
        echo "$_msg_not_available"
    fi
}

_disk_usage()
{
    # report the about of used vs total space for a mounted volume, and the percentage of free space remaining
    # takes 1 argument, which is a path to the disk's mountpoint, as reported by the 'Mounted on' column of 'df'
    # or the filesytstem location of the disk's unmounted block device e.g. /dev/sda2, as reported in the 1st column
    # requires the command 'df', plus the sucpplementary ommand 'bc' to calculate the remaining free space
    _disk="$1"
    _df_output="$(df --output='used,size,pcent' -h $_disk 2> /dev/null)"

    if [ "$_df_output" ]; then
        _df_output="$(echo "$_df_output" | tail -1 | sed -e "s/  */ /g" -e "s/^ *//g")"

        _used="$(echo "$_df_output" | cut -d " " -f1)"
        _total="$(echo "$_df_output" | cut -d " " -f2)"
        _free="$(echo "$_df_output" | cut -d " " -f3)"
        _free="$(echo "100 - ${_free%\%}" | bc)" 

        echo "${_used}/${_total} (${_free}% free)"
    else
        echo "$_msg_not_available"
    fi
}

_disk_health()
{
    # for each real physical disk detected on the system by the command 'lsscsi'
    # check the output of smartctl -x, for the answer to: "SMART overall-health self-assessment test result"
    # and report back any disk status of any real physical disk (that supports SMART protocol) if any of those disks
    # have a test result status anything other than 'PASSED'
    # requires the package 'smartmontools', for the command 'smartctl'
    # requires the package 'lsscsi' for the command 'lsscsi'

    if [ "$(command -v smartctl)" ]; then
        _disks="$(lsscsi | sed -e "s|.*/dev/||g")"

        if ! _try_sudo; then
            echo "$_msg_not_available - sudo smartctl: failed"
            return 1
        fi

        unset _output_msg
        for _disk in $_disks; do
            _disk_health="$($_sudo smartctl --health /dev/${_disk} 2>&1 | grep -i "test result:" | sed -e "s/.*test result: //g")"
            if [ "$_disk_health" ] && [ "$_disk_health" != "PASSED" ]; then
                _disk_name="$(lsscsi | grep "/dev/${_disk}" | sed -e "s/.*ATA *//g" -e "s| */dev/.*||g" -e "s/ [^ ]*$//g")"
                _output_msg="${_output_msg}, ${_disk_name}(${_disk})=${_disk_health}"
            fi
        done

        if [ "$_output_msg" ]; then
            echo "${_output_msg#, }"
        else
            echo "ok"
        fi

    else
        echo "$_msg_not_available"
    fi
}

_disks_worst_temp()
{
    # for each real physical disk detected on the system by the command 'lsscsi'
    # check the output of smartctl -x, for the disks temperature information lines
    # for the current temperature and the maximum recommended temperature
    # if the current temperature is too close to the maximum temperature (default 10c, see below)
    # then print out a CRITICAL: message to warn about the situation
    # otherwise just report ok, along with the current 'worst disk temp' that is closest to that threshold
    # requires the package 'smartmontools', for the command 'smartctl'
    # requires the package 'lsscsi' for the command 'lsscsi'

    if ! _try_sudo; then
        echo "$_msg_not_available - sudo smartctl: failed"
        return 1
    fi

    if [ "$(command -v smartctl)" ]; then
        _disks="$(lsscsi | sed -e "s|.*/dev/||g")"

        # set here the threshold distance in degrees C, to which to warn if the drive is nearing it's
        # own recommended maximum temperature, as read from its own individual smart data
        # for example, if you don't feel comfortable being within 10c, 5c or 0c of that limit
        # _temp_critical_distance_to_reccomended_max=100 # just for testing
        _temp_critical_distance_to_reccomended_max=10
        # _temp_critical_distance_to_reccomended_max=5
        # _temp_critical_distance_to_reccomended_max=3
        # _temp_critical_distance_to_reccomended_max=0

        # to find and report on the current status of the 'thermally worst' disk in your array
        _temp_last_known_most_critical=99999

        unset _output_msg _worst_disk_temp_msg _crit_disk_temp_msg
        for _disk in $_disks; do
            _smartctl_temp_lines="$($_sudo smartctl -x /dev/${_disk} | grep -i temperature)"

            if [ "$_smartctl_temp_lines" ]; then
                _temp_current="$(echo "$_smartctl_temp_lines" | grep "Current Temperature:" | sed -e "s/.*: *//g" -e "s/ .*//g")"
                _temp_max_recommended="$(echo "$_smartctl_temp_lines" | grep -i "recommended Temperature:" | sed -e "s/.*: *//g" -e "s/ .*//g" -e "s|.*/||g")"
                _temp_distance_to_max="$(echo "$_temp_max_recommended - $_temp_current" | bc)"

                if [ $_temp_distance_to_max -lt $_temp_last_known_most_critical ]; then
                    _worst_disk_temp_msg="${_disk}: ${_temp_current}c/${_temp_max_recommended}c"
                fi

                if [ $_temp_distance_to_max -le $_temp_critical_distance_to_reccomended_max ]; then
                    _crit_disk_temp_msg="${_crit_disk_temp_msg}, ${_disk}:${_temp_current}c/${_temp_max_recommended}c"
                fi
            fi
        done

        if [ "$_crit_disk_temp_msg" ]; then
            echo "CRITICAL!!! ${_crit_disk_temp_msg#, }"
        else
            echo "ok (${_worst_disk_temp_msg})"
        fi
    else
        echo "$_msg_not_available"
    fi
}

_zfs_health()
{
    # for each zpool found on the system (that is being reported by the command 'zpool')
    # check it's status with the command 'zpool status', and either report 'ok', if all pools are 
    # in the status 'ONLINE', otherwise report the status of any pools that are not 'ONLINE'
    # and make a best efforts to interperet and to report back the reason for being OFFLINE, DEGRADED, etc. 
    # requires the 'zfs' package(s), for the command 'zpool', and a working zfs filesystem loaded
    unset _zpool_errors _output_line
    if [ "$(command -v zpool)" ]; then

        if ! zpool list -H > /dev/null 2>&1; then
            echo "$_msg_not_available"
            return 1
        fi

        _zpools="$(zpool list -H 2>&1 | cut -f1)"

        for _zpool in $_zpools; do

            _zpool_status="$(zpool status -v $_zpool 2>&1)"
            _state="$(echo "$_zpool_status" | grep "^ *state: " | sed -e "s/^ *state: //g")"
            _error_line="$(echo "$_zpool_status" | grep "^errors: " | sed -e "s/^errors: //g")"

            if [ "$_state" = "ONLINE" ]; then
                _output_line="${_output_line},${_zpool}=${_state}"
            else

                if echo "$_error_line" | grep -i -q "List of errors unavailable: "; then
                    _error_detail="$(echo "$_error_line" | sed -e "s/List of errors unavailable: //g")"

                elif echo "$_error_line" | grep -i -q "detected in"; then
                    _error_detail="$(echo "$_error_line" | sed -e "s/detected in.*$/detected/g")"

                elif echo "$_error_line" | grep -i -q "The following "; then
                    _error_detail="$(echo "$_error_line" | sed -e "s/The following //g" -e "s/:$//g")"

                else
                    _error_detail="$_error_line"
                fi

                _zpool_errors=true
                _output_line="${_output_line},!!!${_zpool}=${_state} (${_error_detail})"
            fi
        done


        if [ "$_zpool_errors" ]; then
            echo "${_output_line#,*}"
        else
            echo "ok"
        fi

    else
        echo "$_msg_not_available"
    fi
}

_systemd_health()
{
    # check the overall health status of systemd services. if there are any failed services
    # then just report how many total failed services there are, otherwise report 'ok'

    # requires systemd, for the command 'systemctl'
    if [ "$(command -v systemctl)" ]; then
        _systemctl_failed="$(systemctl --all --state=failed 2> /dev/null)"
        _num_units_failed="$(echo "$_systemctl_failed" | grep "loaded units listed" | cut -d" " -f1)"
        if [ $_num_units_failed -gt 0 ]; then
            echo "${_num_units_failed} failed services"

        else
            echo "ok"
        fi
    else
        echo "$_msg_not_available"
    fi
}

_systemctl_status()
{
    # check up on the current status of the specified systemd service. takes 1 argument from the commandline
    # which is the name of the systemd service to check, for example "docker.service"

    # requires systemd, for the command 'systemctl'
    if [ "$(command -v systemctl)" ]; then
        _systemctl_service="$1"
        _systemctl_output="$(systemctl status ${_systemctl_service})"
        _service_status="$(echo "$_systemctl_output" | grep "Active: " | sed -e "s/^.*Active: //g" -e "s/ .*//g")"

        if [ "$_service_status" != "active" ]; then
            echo "$_service_status"
        else
            echo "ok"
        fi
    else
        echo "$_msg_not_available"
    fi
}

_docker_status()
{
    # report back the current status of the docker daemon, if it is running or not
    # and the current number of running containers, vs the total number of running + stopped containers

    # requires systemd, for the command 'systemctl', and
    # requires the package 'docker-ce', for the command 'docker'
    if [ "$(command -v docker)" ]; then

        _docker_service_status="$(_systemctl_status "docker.service")"

        if [ "$_docker_service_status" = "ok" ]; then
            _docker_info="$(docker info 2>&1)"
            _num_containers_running="$(echo "$_docker_info" | grep -m1 "Running: " | sed -e "s/.*Running: //g")"
            _num_containers_total="$(echo "$_docker_info" | grep -m1 "Containers: " | sed -e "s/.*Containers: //g")"
            _server_version="$(echo "$_docker_info" | grep -m1 "Server Version: " | sed -e "s/.*Server Version: //g")"

            echo "ok (v${_server_version}), ${_num_containers_running}/${_num_containers_total} running"
        else
            echo "$_docker_service_status"
        fi

    else
        echo "$_msg_not_available"
    fi
}

_socket_send()
{
    # requires the package "socat" for the command "socat"
    # client - send 1 command, get the response, then exit
    _socket_file="$1"
    shift

    _pid_file="/tmp/socat.pid"

    # unfortunately this next line is taken from an example which requires bash
    # don't know how to achieve the same objective in bourne sh shell / dash
    (cat <( echo "$@" ) - | socat - "UNIX:${_socket_file}" & echo $! >&3) 3>$_pid_file | \
    while read -r _reply; do

        echo "$_reply"
        kill $(cat "$_pid_file")
        break
    done
}

_serve_request()
{
    while true; do
        read -r _line

        if [ "$_line" = "socket-listen" ] || [ "$_line" = "socket-send" ]; then
            echo "error: nesting sockets is not permitted"
        else
            _parse_args $_line
        fi
    done
}

_socket_listen()
{
    # this funny trick is using read twice over, to return subsequent output back into the front of the pipe
    # solution: https://stackoverflow.com/a/43332/287510

    # open up socat and wait
    _socket_file="$1"
    read | { socat "UNIX-LISTEN:${_socket_file},fork" - | _serve_request; } >/dev/fd/0
}


_parse_args()
{
    unset _arg _arg_disk_usage _socket
    if [ ! "$1" ]; then
        _no_args=true
    fi

    while [ "$1" ]; do
        _arg="$1"

        if [ "$_send_socket" ]; then
            _socket_send "$_send_socket" "$@"
            exit 0

        else
            case $_arg in

                platform)           _platform;;
                kernel)             _kernel;;
                os)                 _os;;
                uptime)             _uptime;;
                cpu-cores)          _cpu_cores;;
                cpu-usage)          _cpu_usage;;
                cpu-temp)           _cpu_temp;;
                fan-spin)           _arg_fan_spin=true;;
                memory-usage)       _memory_usage;;
                dmesg-health)       _dmesg_health;;
                disks)              _disks;;
                disks-worst-temp)   _disks_worst_temp;;
                disk-usage)         _arg_disk_usage=true;;
                disk-health)        _disk_health;;
                zfs-health)         _zfs_health;;
                systemd-health)     _systemd_health;;
                systemctl-status)   _arg_systemctl_status=true;;
                docker-status)      _docker_status;;
                socket-listen)      _arg_socket_listen=true;;
                socket-send)        _arg_socket_send=true;;

                --debug|-d)                  _debug=true; set -x;;
                --help|-h)                   _help=true ;;

                *)
                if [ "$_arg_fan_spin" ]; then
                    _fans="$_arg"

                elif [ "$_arg_disk_usage" ]; then
                    _disk="$_arg"

                elif [ "$_arg_systemctl_status" ]; then
                    _systemctl_service="$_arg"

                elif [ "$_arg_socket_listen" ]; then
                    _listen_socket="$_arg"
                    _socket_listen "$_listen_socket"
                    return 0

                elif [ "$_arg_socket_send" ]; then
                    _send_socket="$_arg"

                else
                    warn "unrecognized argument: \"$_arg\""
                    _cat_help
                    exit 1
                fi
                ;;
            esac
        fi


        shift
    done

    if [ "$_help" ] || [ "$_no_args" ]; then
        _cat_help
        exit 0
    fi


    if [ "$_arg_fan_spin" ]; then
        if [ ! "$_fans" ]; then
            err 1 "no fans specified"
        else
            _fan_spin "$_fans"
        fi
    fi

    if [ "$_arg_disk_usage" ]; then
        if [ ! "$_disk" ]; then
            _disk="/"
        fi
        _disk_usage "$_disk"
    fi

    if [ "$_arg_systemctl_status" ]; then
        if [ ! "$_systemctl_service" ]; then
            err 1 "no systemd service specified"
        else
            _systemctl_status "$_systemctl_service"
        fi
    fi
}










# =========================
# begin:

_parse_args "$@"