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
|
#!/bin/bash
. $(dirname $0)/../include.rc
. $(dirname $0)/../volume.rc
. $(dirname $0)/../nfs.rc
. $(dirname $0)/../fileio.rc
#G_TESTDEF_TEST_STATUS_CENTOS6=NFS_TEST
cleanup;
QDD=$(dirname $0)/quota
# compile the test write program and run it
build_tester $(dirname $0)/quota.c -o $QDD
TESTS_EXPECTED_IN_LOOP=16
TEST glusterd
TEST pidof glusterd
TEST $CLI volume info;
TEST $CLI volume create $V0 $H0:$B0/brick1;
EXPECT 'Created' volinfo_field $V0 'Status';
TEST $CLI volume set $V0 nfs.disable false
# The test makes use of inode-lru-limit to hit a scenario, where we
# find an inode whose ancestry is not there. Following is the
# hypothesis (which is confirmed by seeing logs indicating that
# codepath has been executed, but not through a good understanding of
# NFS internals).
# At the end of an fop, the reference count of an inode would be
# zero. The inode (and its ancestry) persists in memory only
# because of non-zero lookup count. These looked up inodes are put
# in an lru queue of size 1 (here). So, there can be at most one
# such inode in memory.
# NFS Server makes use of anonymous fds. So, if it cannot find
# valid fd, it does a nameless lookup. This gives us an inode
# whose ancestry is NULL. When a write happens on this inode,
# quota-enforcer/marker finds a NULL ancestry and asks
# storage/posix to build it.
TEST $CLI volume set $V0 network.inode-lru-limit 1
TEST $CLI volume set $V0 performance.nfs.write-behind off
TEST $CLI volume start $V0;
EXPECT 'Started' volinfo_field $V0 'Status';
TEST $CLI volume quota $V0 enable
TEST $CLI volume quota $V0 limit-usage / 1
TEST $CLI volume quota $V0 soft-timeout 0
TEST $CLI volume quota $V0 hard-timeout 0
EXPECT_WITHIN $NFS_EXPORT_TIMEOUT "1" is_nfs_export_available;
TEST mount_nfs $H0:/$V0 $N0 noac,soft,nolock,vers=3;
deep=/0/1/2/3/4/5/6/7/8/9
TEST mkdir -p $N0/$deep
TEST touch $N0/$deep/file1 $N0/$deep/file2 $N0/$deep/file3 $N0/$deep/file4
TEST fd_open 3 'w' "$N0/$deep/file1"
TEST fd_open 4 'w' "$N0/$deep/file2"
TEST fd_open 5 'w' "$N0/$deep/file3"
TEST fd_open 6 'w' "$N0/$deep/file4"
# consume all quota
echo "Hello" > $N0/$deep/new_file_1
echo "World" >> $N0/$deep/new_file_1
echo 1 >> $N0/$deep/new_file_1
echo 2 >> $N0/$deep/new_file_1
# Try to create a 1M file which should fail
TEST ! $QDD $N0/$deep/new_file_2 256 4
# At the end of each fop in server, reference count of the
# inode associated with each of the file above drops to zero and hence
# put into lru queue. Since lru-limit is set to 1, an fop next file
# will displace the current inode from itable. This will ensure that
# when writes happens on same fd, fd resolution results in
# nameless lookup from server and quota_writev encounters an fd
# associated with an inode whose parent is not present in itable.
for j in $(seq 1 2); do
for i in $(seq 3 6); do
# failing writes indicate that we are enforcing quota set on /
# even with anonymous fds.
TEST_IN_LOOP ! fd_write $i "content"
TEST_IN_LOOP sync
done
done
exec 3>&-
exec 4>&-
exec 5>&-
exec 6>&-
$CLI volume statedump $V0 all
EXPECT_WITHIN $UMOUNT_TIMEOUT "Y" force_umount $N0
# This is ugly, but there seems to be a latent race between other actions and
# stopping the volume. The visible symptom is that "umount -l" (run from
# gf_umount_lazy in glusterd) hangs. This happens pretty consistently with the
# new mem-pool code, though it's not really anything to do with memory pools -
# just with changed timing. Adding the sleep here makes it work consistently.
#
# If anyone else wants to debug the race condition, feel free.
sleep 3
TEST $CLI volume stop $V0
rm -f $QDD
cleanup;
#G_TESTDEF_TEST_STATUS_NETBSD7=BAD_TEST,BUG=000000
|
