summaryrefslogtreecommitdiffstats
path: root/extras/stop-all-gluster-processes.sh
blob: 710aaf5fd3cc17a32645168083680a9c0c785bdc (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
#!/bin/bash
#
# Kill all the processes/services except glusterd
#
# Usage: ./extras/stop-all-gluster-processes.sh [-g] [-h]
#    options:
#    -g  Terminate in graceful mode
#    -h  Show this message, then exit
#
# eg:
#  1. ./extras/stop-all-gluster-processes.sh
#  2. ./extras/stop-all-gluster-processes.sh -g
#
# By default, this script executes in force mode, i.e. all of brick, gsyncd
# and other glustershd services/processes are killed without checking for
# ongoing tasks such as geo-rep, self-heal, rebalance and etc. which may lead
# to inconsistency after the node is brought back.
#
# On specifying '-g' option this script works in graceful mode, to maintain
# data consistency the script fails with a valid exit code incase if any of
# the gluster processes are busy in doing their jobs.
#
# The author of page [1] proposes user-defined exit codes to the range 64 - 113
# Find the better explanation behind the choice in the link
#
# The exit code returned by stop-all-gluster-processes.sh:
#   0       No errors/Success
#   64      Rebalance is in progress
#   65      Self-Heal is in progress
#   66      Tier daemon running on this node
#   127     option not found
#
# [1] http://www.tldp.org/LDP/abs/html/exitcodes.html


# global
errors=0

# find the mounts and return their pids
get_mount_pids()
{
    local opts
    local pid

    for opts in $(grep -w fuse.glusterfs /proc/mounts| awk '{print $1":/"$2}');
    do
        IFS=' ' read -r -a volinfo <<< $(echo "${opts}" | sed 's/:\// /g')
        pid+="$(ps -Ao pid,args | grep -w "volfile-server=${volinfo[0]}" |
                grep -w "volfile-id=/${volinfo[1]}" | grep -w "${volinfo[2]}" |
                awk '{print $1}') "
    done
    echo "${pid}"
}

# handle mount processes i.e. 'glusterfs'
kill_mounts()
{
    local signal=${1}
    local pid

    for pid in $(get_mount_pids);
    do
        echo "sending SIG${signal} to mount process with pid: ${pid}";
        kill -${signal} ${pid};
    done
}

# handle brick processes and node services
kill_bricks_and_services()
{
    local signal=${1}
    local pidfile
    local pid

    for pidfile in $(find /var/run/gluster/ -name '*.pid');
    do
        local pid=$(cat ${pidfile});
        echo "sending SIG${signal} to pid: ${pid}";
        kill -${signal} ${pid};
    done
}

# for geo-replication, only 'monitor' has pid file written, other
# processes are not having a pid file, so get it through 'ps' and
# handle these processes
kill_georep_gsync()
{
    local signal=${1}

    # FIXME: add strick/better check
    local gsyncpid=$(ps -Ao pid,args | grep gluster | grep gsync |
                     awk '{print $1}');
    if [ -n "${gsyncpid}" ]
    then
        echo "sending SIG${signal} to geo-rep gsync process ${gsyncpid}";
        kill -${signal} ${gsyncpid} || errors=$((${errors} + 1));
    fi
}

# check if all processes are ready to die
check_background_tasks()
{
    volumes=$(gluster vol list)
    quit=0
    for volname in ${volumes};
    do
        # tiering
        if [[ $(gluster volume tier ${volname} status 2> /dev/null |
                grep "localhost" | grep -c "in progress") -gt 0 ]]
        then
            quit=66
            break;
        fi

        # rebalance
        if [[ $(gluster volume rebalance ${volname} status 2> /dev/null |
                grep -c "in progress") -gt 0 ]]
        then
            quit=64
            break;
        fi

        # self heal
        if [[ $(gluster volume heal ${volname} info | grep "Number of entries" |
                awk '{ sum+=$4} END {print sum}') -gt 0 ]];
        then
            quit=65
            break;
        fi

        # geo-rep, snapshot and quota doesn't need grace checks,
        # as they ensures the consistancy on force kills
    done

    echo ${quit}
}

usage()
{
    cat <<EOM
Usage: $0 [-g] [-h]
    options:
    -g  Terminate in graceful mode
    -h  Show this message, then exit

eg:
 1. $0
 2. $0 -g
EOM
}

main()
{
    while getopts "gh" opt; do
        case $opt in
            g)
                # graceful mode
                quit=$(check_background_tasks)
                if [[ ${quit} -ne 0 ]]
                then
                    exit ${quit};
                fi
                # else safe to kill
                ;;
            h)
                usage
                exit 0;
                ;;
            *)
                usage
                exit 127;
                ;;
        esac
    done
    # remove all the options that have been parsed by getopts
    shift $((OPTIND-1))

    kill_mounts TERM
    kill_georep_gsync TERM
    kill_bricks_and_services TERM

    sleep 5;
    echo ""

    # still not Terminated? let's pass SIGKILL
    kill_mounts KILL
    kill_georep_gsync KILL
    kill_bricks_and_services KILL

    exit ${errors};
}

main "$@"