root / qa / qa_daemon.py @ cec9845c
History | View | Annotate | Download (4.2 kB)
1 |
# Copyright (C) 2007 Google Inc.
|
---|---|
2 |
#
|
3 |
# This program is free software; you can redistribute it and/or modify
|
4 |
# it under the terms of the GNU General Public License as published by
|
5 |
# the Free Software Foundation; either version 2 of the License, or
|
6 |
# (at your option) any later version.
|
7 |
#
|
8 |
# This program is distributed in the hope that it will be useful, but
|
9 |
# WITHOUT ANY WARRANTY; without even the implied warranty of
|
10 |
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
11 |
# General Public License for more details.
|
12 |
#
|
13 |
# You should have received a copy of the GNU General Public License
|
14 |
# along with this program; if not, write to the Free Software
|
15 |
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
|
16 |
# 02110-1301, USA.
|
17 |
|
18 |
|
19 |
"""Daemon related QA tests.
|
20 |
|
21 |
"""
|
22 |
|
23 |
import time |
24 |
import subprocess |
25 |
|
26 |
from ganeti import utils |
27 |
from ganeti import constants |
28 |
|
29 |
import qa_config |
30 |
import qa_utils |
31 |
import qa_error |
32 |
|
33 |
from qa_utils import AssertEqual, StartSSH |
34 |
|
35 |
|
36 |
def _ResolveInstanceName(instance): |
37 |
"""Gets the full Xen name of an instance.
|
38 |
|
39 |
"""
|
40 |
master = qa_config.GetMasterNode() |
41 |
|
42 |
info_cmd = utils.ShellQuoteArgs(['gnt-instance', 'info', instance['name']]) |
43 |
sed_cmd = utils.ShellQuoteArgs(['sed', '-n', '-e', 's/^Instance name: *//p']) |
44 |
|
45 |
cmd = '%s | %s' % (info_cmd, sed_cmd)
|
46 |
ssh_cmd = qa_utils.GetSSHCommand(master['primary'], cmd)
|
47 |
p = subprocess.Popen(ssh_cmd, shell=False, stdout=subprocess.PIPE)
|
48 |
AssertEqual(p.wait(), 0)
|
49 |
|
50 |
return p.stdout.read().strip()
|
51 |
|
52 |
|
53 |
def _InstanceRunning(node, name): |
54 |
"""Checks whether an instance is running.
|
55 |
|
56 |
Args:
|
57 |
node: Node the instance runs on
|
58 |
name: Full name of Xen instance
|
59 |
"""
|
60 |
cmd = utils.ShellQuoteArgs(['xm', 'list', name]) + ' >/dev/null' |
61 |
ret = StartSSH(node['primary'], cmd).wait()
|
62 |
return ret == 0 |
63 |
|
64 |
|
65 |
def _XmShutdownInstance(node, name): |
66 |
"""Shuts down instance using "xm" and waits for completion.
|
67 |
|
68 |
Args:
|
69 |
node: Node the instance runs on
|
70 |
name: Full name of Xen instance
|
71 |
"""
|
72 |
master = qa_config.GetMasterNode() |
73 |
|
74 |
cmd = ['xm', 'shutdown', name] |
75 |
AssertEqual(StartSSH(master['primary'],
|
76 |
utils.ShellQuoteArgs(cmd)).wait(), 0)
|
77 |
|
78 |
# Wait up to a minute
|
79 |
end = time.time() + 60
|
80 |
while time.time() <= end:
|
81 |
if not _InstanceRunning(node, name): |
82 |
break
|
83 |
time.sleep(5)
|
84 |
else:
|
85 |
raise qa_error.Error("xm shutdown failed") |
86 |
|
87 |
|
88 |
def _ResetWatcherDaemon(node): |
89 |
"""Removes the watcher daemon's state file.
|
90 |
|
91 |
Args:
|
92 |
node: Node to be reset
|
93 |
"""
|
94 |
cmd = ['rm', '-f', constants.WATCHER_STATEFILE] |
95 |
AssertEqual(StartSSH(node['primary'],
|
96 |
utils.ShellQuoteArgs(cmd)).wait(), 0)
|
97 |
|
98 |
|
99 |
def TestInstanceAutomaticRestart(node, instance): |
100 |
"""Test automatic restart of instance by ganeti-watcher.
|
101 |
|
102 |
Note: takes up to 6 minutes to complete.
|
103 |
"""
|
104 |
master = qa_config.GetMasterNode() |
105 |
inst_name = _ResolveInstanceName(instance) |
106 |
|
107 |
_ResetWatcherDaemon(node) |
108 |
_XmShutdownInstance(node, inst_name) |
109 |
|
110 |
# Give it a bit more than five minutes to start again
|
111 |
restart_at = time.time() + 330
|
112 |
|
113 |
# Wait until it's running again
|
114 |
while time.time() <= restart_at:
|
115 |
if _InstanceRunning(node, inst_name):
|
116 |
break
|
117 |
time.sleep(15)
|
118 |
else:
|
119 |
raise qa_error.Error("Daemon didn't restart instance in time") |
120 |
|
121 |
cmd = ['gnt-instance', 'info', inst_name] |
122 |
AssertEqual(StartSSH(master['primary'],
|
123 |
utils.ShellQuoteArgs(cmd)).wait(), 0)
|
124 |
|
125 |
|
126 |
def TestInstanceConsecutiveFailures(node, instance): |
127 |
"""Test five consecutive instance failures.
|
128 |
|
129 |
Note: takes at least 35 minutes to complete.
|
130 |
"""
|
131 |
master = qa_config.GetMasterNode() |
132 |
inst_name = _ResolveInstanceName(instance) |
133 |
|
134 |
_ResetWatcherDaemon(node) |
135 |
_XmShutdownInstance(node, inst_name) |
136 |
|
137 |
# Do shutdowns for 30 minutes
|
138 |
finished_at = time.time() + (35 * 60) |
139 |
|
140 |
while time.time() <= finished_at:
|
141 |
if _InstanceRunning(node, inst_name):
|
142 |
_XmShutdownInstance(node, inst_name) |
143 |
time.sleep(30)
|
144 |
|
145 |
# Check for some time whether the instance doesn't start again
|
146 |
check_until = time.time() + 330
|
147 |
while time.time() <= check_until:
|
148 |
if _InstanceRunning(node, inst_name):
|
149 |
raise qa_error.Error("Instance started when it shouldn't") |
150 |
time.sleep(30)
|
151 |
|
152 |
cmd = ['gnt-instance', 'info', inst_name] |
153 |
AssertEqual(StartSSH(master['primary'],
|
154 |
utils.ShellQuoteArgs(cmd)).wait(), 0)
|