blob: 13b5218372219536c8b475a0b6275c0c000867f5 (
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
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
|
#!/bin/bash
. $(dirname $0)/../../include.rc
. $(dirname $0)/../../volume.rc
. $(dirname $0)/../../fileio.rc
cleanup;
TEST glusterd;
TEST pidof glusterd
TEST $CLI volume create $V0 replica 3 $H0:$B0/${V0}{0,1,2}
TEST $CLI volume set $V0 performance.write-behind off
TEST $CLI volume set $V0 performance.open-behind off
TEST $CLI volume set $V0 client.strict-locks on
TEST $CLI volume heal $V0 disable
TEST $CLI volume start $V0
EXPECT 'Started' volinfo_field $V0 'Status';
TEST $GFS --volfile-id=/$V0 --volfile-server=$H0 $M0
TEST $GFS --volfile-id=/$V0 --volfile-server=$H0 $M1
TEST touch $M0/a
# Kill one brick and take lock on the fd and do a write.
TEST kill_brick $V0 $H0 $B0/${V0}0
EXPECT_WITHIN $PROCESS_DOWN_TIMEOUT "0" afr_child_up_status_meta $M0 $V0-replicate-0 0
TEST fd1=`fd_available`
TEST fd_open $fd1 'rw' $M0/a
TEST flock -x $fd1
TEST fd_write $fd1 "data-1"
# Restart the brick and then write. Now fd should not get re-opened but write
# should still succeed as there were no quorum disconnects.
TEST $CLI volume start $V0 force
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^1$" brick_up_status $V0 $H0 $B0/${V0}0
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status_meta $M0 $V0-replicate-0 0
TEST fd_write $fd1 "data-2"
EXPECT "" cat $B0/${V0}0/a
EXPECT "data-2" cat $B0/${V0}1/a
EXPECT "data-2" cat $B0/${V0}2/a
# Check there is no fd opened on the 1st brick by checking for the gfid inside
# /proc/pid-of-brick/fd/ directory
gfid_a=$(gf_get_gfid_xattr $B0/${V0}0/a)
gfid_str_a=$(gf_gfid_xattr_to_str $gfid_a)
EXPECT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
TEST fd2=`fd_available`
TEST fd_open $fd2 'rw' $M1/a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^2$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT "^2$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
# Kill 2nd brick and try writing to the file. The write should fail due to
# quorum failure.
TEST kill_brick $V0 $H0 $B0/${V0}1
EXPECT_WITHIN $PROCESS_DOWN_TIMEOUT "0" afr_child_up_status_meta $M0 $V0-replicate-0 1
TEST ! fd_write $fd1 "data-3"
TEST ! fd_cat $fd1
# Restart the bricks and try writing to the file. This should fail as two bricks
# which were down previously, will return EBADFD now.
TEST $CLI volume start $V0 force
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^1$" brick_up_status $V0 $H0 $B0/${V0}1
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status_meta $M0 $V0-replicate-0 1
TEST ! fd_write $fd1 "data-4"
TEST ! fd_cat $fd1
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT "^2$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
# Enable heal and check the files will have same content on all the bricks after
# the heal is completed.
EXPECT_WITHIN $HEAL_TIMEOUT "^2$" get_pending_heal_count $V0
TEST $CLI volume heal $V0 enable
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "Y" glustershd_up_status
EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0
EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1
EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 2
TEST $CLI volume heal $V0
EXPECT_WITHIN $HEAL_TIMEOUT "^0$" get_pending_heal_count $V0
EXPECT "data-4" cat $B0/${V0}0/a
EXPECT "data-4" cat $B0/${V0}1/a
EXPECT "data-4" cat $B0/${V0}2/a
TEST $CLI volume heal $V0 disable
# Try writing to the file again on the same fd, which should fail again, since
# it is not yet re-opened.
TEST ! fd_write $fd1 "data-5"
# At this point only one brick will have the lock. Try taking the lock again on
# the bad fd, which should also fail with EBADFD.
# TODO: At the moment quorum failure in lk leads to unlock on the bricks where
# lock succeeds. This will change lock state on 3rd brick, commenting for now
#TEST ! flock -x $fd1
# Kill the only brick that is having lock and try taking lock on another client
# which should succeed.
TEST kill_brick $V0 $H0 $B0/${V0}2
EXPECT_WITHIN $PROCESS_DOWN_TIMEOUT "0" afr_child_up_status_meta $M0 $V0-replicate-0 2
TEST flock -x $fd2
TEST fd_write $fd2 "data-6"
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
# Bring the brick up and try writing & reading on the old fd, which should still
# fail and operations on the 2nd fd should succeed.
TEST $CLI volume start $V0 force
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^1$" brick_up_status $V0 $H0 $B0/${V0}2
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status_meta $M0 $V0-replicate-0 2
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status_meta $M1 $V0-replicate-0 2
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
TEST ! fd_write $fd1 "data-7"
TEST ! fd_cat $fd1
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
TEST fd_cat $fd2
# Close both the fds which will release the locks and then re-open and take lock
# on the old fd. Operations on that fd should succeed afterwards.
TEST fd_close $fd1
TEST fd_close $fd2
EXPECT_WITHIN $REOPEN_TIMEOUT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT_WITHIN $REOPEN_TIMEOUT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT_WITHIN $REOPEN_TIMEOUT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
TEST fd1=`fd_available`
TEST fd_open $fd1 'rw' $M0/a
EXPECT_WITHIN $REOPEN_TIMEOUT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT_WITHIN $REOPEN_TIMEOUT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT_WITHIN $REOPEN_TIMEOUT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
TEST flock -x $fd1
TEST fd_write $fd1 "data-8"
TEST fd_cat $fd1
EXPECT "data-8" head -n 1 $B0/${V0}0/a
EXPECT "data-8" head -n 1 $B0/${V0}1/a
EXPECT "data-8" head -n 1 $B0/${V0}2/a
TEST fd_close $fd1
EXPECT_WITHIN $REOPEN_TIMEOUT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT_WITHIN $REOPEN_TIMEOUT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT_WITHIN $REOPEN_TIMEOUT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
# Heal the volume
TEST $CLI volume heal $V0 enable
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "Y" glustershd_up_status
EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0
EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1
EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 2
TEST $CLI volume heal $V0
EXPECT_WITHIN $HEAL_TIMEOUT "^0$" get_pending_heal_count $V0
TEST $CLI volume heal $V0 disable
# Kill one brick and open a fd.
TEST kill_brick $V0 $H0 $B0/${V0}0
EXPECT_WITHIN $PROCESS_DOWN_TIMEOUT "0" afr_child_up_status_meta $M0 $V0-replicate-0 0
TEST fd1=`fd_available`
TEST fd_open $fd1 'rw' $M0/a
EXPECT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
# Restart the brick and then write. Now fd should get re-opened and write should
# succeed on the previously down brick as well since there are no locks held on
# any of the bricks.
TEST $CLI volume start $V0 force
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^1$" brick_up_status $V0 $H0 $B0/${V0}0
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status_meta $M0 $V0-replicate-0 0
TEST fd_write $fd1 "data-10"
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "data-10" head -n 1 $B0/${V0}0/a
EXPECT "data-10" head -n 1 $B0/${V0}1/a
EXPECT "data-10" head -n 1 $B0/${V0}2/a
TEST fd_close $fd1
# Kill one brick, open and take lock on a fd.
TEST kill_brick $V0 $H0 $B0/${V0}0
EXPECT_WITHIN $PROCESS_DOWN_TIMEOUT "0" afr_child_up_status_meta $M0 $V0-replicate-0 0
TEST fd1=`fd_available`
TEST fd_open $fd1 'rw' $M0/a
TEST flock -x $fd1
EXPECT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
# Kill & restart another brick so that it will return EBADFD
TEST kill_brick $V0 $H0 $B0/${V0}1
EXPECT_WITHIN $PROCESS_DOWN_TIMEOUT "0" brick_up_status $V0 $H0 $B0/${V0}1
# Restart the bricks and then write. Now fd should not get re-opened since lock
# is still held on one brick and write should also fail as there is no quorum.
TEST $CLI volume start $V0 force
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^1$" brick_up_status $V0 $H0 $B0/${V0}0
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^1$" brick_up_status $V0 $H0 $B0/${V0}1
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status_meta $M0 $V0-replicate-0 0
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status_meta $M0 $V0-replicate-0 1
TEST ! fd_write $fd1 "data-11"
EXPECT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}0 $gfid_str_a
EXPECT "^0$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}1 $gfid_str_a
EXPECT "^1$" gf_open_file_count_in_brick $V0 $H0 $B0/${V0}2 $gfid_str_a
EXPECT "data-10" head -n 1 $B0/${V0}0/a
EXPECT "data-10" head -n 1 $B0/${V0}1/a
EXPECT "data-11" head -n 1 $B0/${V0}2/a
TEST fd_close $fd1
cleanup
|
