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
|
#!/bin/sh
# PCP QA Test No. 1286
# Exercise the BCC PMDA netproc hits module - install, remove and values.
#
# Copyright (c) 2020 Red Hat.
#
seq=`basename $0`
echo "QA output created by $seq"
. ./common.bcc
_pmdabcc_check
_pmdabcc_require_kernel_version 4 6
which ss >/dev/null 2>&1 || _notrun "No ss binary installed"
status=1 # failure is the default!
signal=$PCP_BINADM_DIR/pmsignal
# Some combinations of distro and kernel are known to be "bad"
# for some BPF modules
#
kernel=`uname -r`
echo "$kernel" >>$seq_full
case `admin/whatami`
in
*Fedora\ 32\ *)
_notrun "bcc netproc module will not compile on Fedora 32"
;;
*Ubuntu\ 20.04\ *)
# works on 5.4 kernel, not on 5.8.0-1033-azure
#
case "$kernel"
in
5.8.0-*-azure)
_notrun "bcc netproc module will not compile on Ubuntu kernel $kernel"
;;
esac
;;
esac
# The netproc BCC module uses the lookup_or_try_init() method that was
# introduced in bcc v0.12.0 (released 11 Dec 2019), so determine the
# bcc version and _notrun as appropriate.
#
# Like most things in Pythonland, it would appear there are many (more
# that 6 according to google) ways to determine the version of a module,
# but none of them are guaranteed to work! The recipe below seems robust
# for the bcc module.
#
cat <<'End-of-File' >$tmp.py
import bcc
print(bcc.__version__)
End-of-File
echo "bcc version ..." >>$seq_full
version=`pmpython $tmp.py 2>>$seq_full`
echo "$version" >>$seq_full
case "$version"
in
'')
# failed to extract version (see $seq.full for more info) ...
# but punt on running the test
;;
0.0.*|0.1[0-1].*)
_notrun "need version >= 0.12.0 (not $version) for bcc Python module"
;;
esac
_pid_filter()
{
sed \
-e "s/0*$1/SERVERPID/g" \
-e "s/0*$2/CLIENTPID/g" \
-e '/inst \[[0-9]/d' \
#end
}
cat <<EOF >$tmp.conf
[pmda]
modules = netproc
prefix = bcc.
[netproc]
module = netproc
cluster = 40
pmda_indom_cache = False
remove_stopped_processes = False
EOF
_pmdabcc_try_compile $tmp.conf
_prepare_pmda bcc
trap "_pmdabcc_cleanup; exit \$status" 0 1 2 3 15
_stop_auto_restart pmcd
# real QA test starts here
_pmdabcc_install <$tmp.conf
_pmdabcc_wait_for_metric
# Generate system activity for the BCC netproc module
$python src/bcc_netproc.py server >> "$seq_full" &
server_pid=$!
# wait for server to listen on specified port
for i in `seq 1 10`; do ss -lnt | grep -q :1234 && break; sleep 1; done
if [ $i -ge 10 ]; then
echo "Server didn't open tcp/1234 in time, test failed"
exit 1
fi
$python src/bcc_netproc.py client >> "$seq_full" &
client_pid=$!
wait ${client_pid}
wait ${server_pid}
echo "server PID: ${server_pid}" >> "$seq_full"
echo "client PID: ${client_pid}" >> "$seq_full"
for metric in bcc.proc.net.tcp.send.calls \
bcc.proc.net.tcp.send.bytes \
bcc.proc.net.tcp.recv.calls \
bcc.proc.net.tcp.recv.bytes \
bcc.proc.net.udp.send.calls \
bcc.proc.net.udp.send.bytes \
bcc.proc.net.udp.recv.calls \
bcc.proc.net.udp.recv.bytes
do
echo && echo && echo "=== report metric values for $metric ==="
pminfo -dfmtT $metric 2>&1 | tee -a $seq_full \
| _pid_filter ${server_pid} ${client_pid} \
| sort # sort to fix the non deterministic instance order
done
_pmdabcc_remove
status=0
exit
|
