root / tools / burnin @ 6605411d
History | View | Annotate | Download (17.8 kB)
1 |
#!/usr/bin/python |
---|---|
2 |
# |
3 |
|
4 |
# Copyright (C) 2006, 2007 Google Inc. |
5 |
# |
6 |
# This program is free software; you can redistribute it and/or modify |
7 |
# it under the terms of the GNU General Public License as published by |
8 |
# the Free Software Foundation; either version 2 of the License, or |
9 |
# (at your option) any later version. |
10 |
# |
11 |
# This program is distributed in the hope that it will be useful, but |
12 |
# WITHOUT ANY WARRANTY; without even the implied warranty of |
13 |
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
14 |
# General Public License for more details. |
15 |
# |
16 |
# You should have received a copy of the GNU General Public License |
17 |
# along with this program; if not, write to the Free Software |
18 |
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA |
19 |
# 02110-1301, USA. |
20 |
|
21 |
|
22 |
"""Burnin program |
23 |
|
24 |
""" |
25 |
|
26 |
import os |
27 |
import sys |
28 |
import optparse |
29 |
import time |
30 |
from itertools import izip, islice, cycle |
31 |
from cStringIO import StringIO |
32 |
|
33 |
from ganeti import opcodes |
34 |
from ganeti import mcpu |
35 |
from ganeti import constants |
36 |
from ganeti import cli |
37 |
from ganeti import logger |
38 |
from ganeti import errors |
39 |
from ganeti import utils |
40 |
|
41 |
|
42 |
USAGE = ("\tburnin -o OS_NAME [options...] instance_name ...") |
43 |
|
44 |
|
45 |
def Usage(): |
46 |
"""Shows program usage information and exits the program.""" |
47 |
|
48 |
print >> sys.stderr, "Usage:" |
49 |
print >> sys.stderr, USAGE |
50 |
sys.exit(2) |
51 |
|
52 |
|
53 |
def Log(msg): |
54 |
"""Simple function that prints out its argument. |
55 |
|
56 |
""" |
57 |
print msg |
58 |
sys.stdout.flush() |
59 |
|
60 |
|
61 |
class Burner(object): |
62 |
"""Burner class.""" |
63 |
|
64 |
def __init__(self): |
65 |
"""Constructor.""" |
66 |
logger.SetupLogging(constants.LOG_BURNIN, debug=False, stderr_logging=True) |
67 |
self._feed_buf = StringIO() |
68 |
self.nodes = [] |
69 |
self.instances = [] |
70 |
self.to_rem = [] |
71 |
self.opts = None |
72 |
self.cl = cli.GetClient() |
73 |
self.ParseOptions() |
74 |
self.GetState() |
75 |
|
76 |
def ClearFeedbackBuf(self): |
77 |
"""Clear the feedback buffer.""" |
78 |
self._feed_buf.truncate(0) |
79 |
|
80 |
def GetFeedbackBuf(self): |
81 |
"""Return the contents of the buffer.""" |
82 |
return self._feed_buf.getvalue() |
83 |
|
84 |
def Feedback(self, msg): |
85 |
"""Acumulate feedback in our buffer.""" |
86 |
self._feed_buf.write("%s %s\n" % (time.ctime(utils.MergeTime(msg[0])), |
87 |
msg[2])) |
88 |
if self.opts.verbose: |
89 |
Log(msg) |
90 |
|
91 |
def ExecOp(self, op): |
92 |
"""Execute an opcode and manage the exec buffer.""" |
93 |
self.ClearFeedbackBuf() |
94 |
return cli.SubmitOpCode(op, feedback_fn=self.Feedback, cl=self.cl) |
95 |
|
96 |
def ExecJobSet(self, jobs): |
97 |
"""Execute a set of jobs and return once all are done. |
98 |
|
99 |
The method will return the list of results, if all jobs are |
100 |
successfull. Otherwise, OpExecError will be raised from within |
101 |
cli.py. |
102 |
|
103 |
""" |
104 |
self.ClearFeedbackBuf() |
105 |
job_ids = [cli.SendJob(job, cl=self.cl) for job in jobs] |
106 |
Log("- Submitted job IDs %s" % ", ".join(job_ids)) |
107 |
results = [] |
108 |
for jid in job_ids: |
109 |
Log("- Waiting for job %s" % jid) |
110 |
results.append(cli.PollJob(jid, cl=self.cl, feedback_fn=self.Feedback)) |
111 |
|
112 |
return results |
113 |
|
114 |
def ParseOptions(self): |
115 |
"""Parses the command line options. |
116 |
|
117 |
In case of command line errors, it will show the usage and exit the |
118 |
program. |
119 |
|
120 |
""" |
121 |
|
122 |
parser = optparse.OptionParser(usage="\n%s" % USAGE, |
123 |
version="%%prog (ganeti) %s" % |
124 |
constants.RELEASE_VERSION, |
125 |
option_class=cli.CliOption) |
126 |
|
127 |
parser.add_option("-o", "--os", dest="os", default=None, |
128 |
help="OS to use during burnin", |
129 |
metavar="<OS>") |
130 |
parser.add_option("--os-size", dest="os_size", help="Disk size", |
131 |
default=4 * 1024, type="unit", metavar="<size>") |
132 |
parser.add_option("--os-growth", dest="sda_growth", help="Disk growth", |
133 |
default=1024, type="unit", metavar="<size>") |
134 |
parser.add_option("--swap-size", dest="swap_size", help="Swap size", |
135 |
default=4 * 1024, type="unit", metavar="<size>") |
136 |
parser.add_option("--swap-growth", dest="sdb_growth", help="Swap growth", |
137 |
default=1024, type="unit", metavar="<size>") |
138 |
parser.add_option("--mem-size", dest="mem_size", help="Memory size", |
139 |
default=128, type="unit", metavar="<size>") |
140 |
parser.add_option("-v", "--verbose", |
141 |
action="store_true", dest="verbose", default=False, |
142 |
help="print command execution messages to stdout") |
143 |
parser.add_option("--no-replace1", dest="do_replace1", |
144 |
help="Skip disk replacement with the same secondary", |
145 |
action="store_false", default=True) |
146 |
parser.add_option("--no-replace2", dest="do_replace2", |
147 |
help="Skip disk replacement with a different secondary", |
148 |
action="store_false", default=True) |
149 |
parser.add_option("--no-failover", dest="do_failover", |
150 |
help="Skip instance failovers", action="store_false", |
151 |
default=True) |
152 |
parser.add_option("--no-importexport", dest="do_importexport", |
153 |
help="Skip instance export/import", action="store_false", |
154 |
default=True) |
155 |
parser.add_option("--no-startstop", dest="do_startstop", |
156 |
help="Skip instance stop/start", action="store_false", |
157 |
default=True) |
158 |
parser.add_option("--rename", dest="rename", default=None, |
159 |
help="Give one unused instance name which is taken" |
160 |
" to start the renaming sequence", |
161 |
metavar="<instance_name>") |
162 |
parser.add_option("-t", "--disk-template", dest="disk_template", |
163 |
choices=("diskless", "file", "plain", "drbd"), |
164 |
default="drbd", |
165 |
help="Disk template (diskless, file, plain or drbd)" |
166 |
" [drbd]") |
167 |
parser.add_option("-n", "--nodes", dest="nodes", default="", |
168 |
help="Comma separated list of nodes to perform" |
169 |
" the burnin on (defaults to all nodes)") |
170 |
parser.add_option("--iallocator", dest="iallocator", |
171 |
default=None, type="string", |
172 |
help="Perform the allocation using an iallocator" |
173 |
" instead of fixed node spread (node restrictions no" |
174 |
" longer apply, therefore -n/--nodes must not be used") |
175 |
parser.add_option("-p", "--parallel", default=False, action="store_true", |
176 |
dest="parallel", |
177 |
help="Enable parallelization of some operations in" |
178 |
" order to speed burnin or to test granular locking") |
179 |
|
180 |
options, args = parser.parse_args() |
181 |
if len(args) < 1 or options.os is None: |
182 |
Usage() |
183 |
|
184 |
supported_disk_templates = (constants.DT_DISKLESS, |
185 |
constants.DT_FILE, |
186 |
constants.DT_PLAIN, |
187 |
constants.DT_DRBD8) |
188 |
if options.disk_template not in supported_disk_templates: |
189 |
Log("Unknown disk template '%s'" % options.disk_template) |
190 |
sys.exit(1) |
191 |
|
192 |
if options.nodes and options.iallocator: |
193 |
Log("Give either the nodes option or the iallocator option, not both") |
194 |
sys.exit(1) |
195 |
|
196 |
self.opts = options |
197 |
self.instances = args |
198 |
self.bep = { |
199 |
constants.BE_MEMORY: options.mem_size, |
200 |
constants.BE_VCPUS: 1, |
201 |
} |
202 |
self.hvp = {} |
203 |
|
204 |
def GetState(self): |
205 |
"""Read the cluster state from the config.""" |
206 |
if self.opts.nodes: |
207 |
names = self.opts.nodes.split(",") |
208 |
else: |
209 |
names = [] |
210 |
try: |
211 |
op = opcodes.OpQueryNodes(output_fields=["name"], names=names) |
212 |
result = self.ExecOp(op) |
213 |
except errors.GenericError, err: |
214 |
err_code, msg = cli.FormatError(err) |
215 |
Log(msg) |
216 |
sys.exit(err_code) |
217 |
self.nodes = [data[0] for data in result] |
218 |
|
219 |
result = self.ExecOp(opcodes.OpDiagnoseOS(output_fields=["name", "valid"], |
220 |
names=[])) |
221 |
|
222 |
if not result: |
223 |
Log("Can't get the OS list") |
224 |
sys.exit(1) |
225 |
|
226 |
# filter non-valid OS-es |
227 |
os_set = [val[0] for val in result if val[1]] |
228 |
|
229 |
if self.opts.os not in os_set: |
230 |
Log("OS '%s' not found" % self.opts.os) |
231 |
sys.exit(1) |
232 |
|
233 |
def CreateInstances(self): |
234 |
"""Create the given instances. |
235 |
|
236 |
""" |
237 |
self.to_rem = [] |
238 |
mytor = izip(cycle(self.nodes), |
239 |
islice(cycle(self.nodes), 1, None), |
240 |
self.instances) |
241 |
jobset = [] |
242 |
|
243 |
for pnode, snode, instance in mytor: |
244 |
if self.opts.iallocator: |
245 |
pnode = snode = None |
246 |
Log("- Add instance %s (iallocator: %s)" % |
247 |
(instance, self.opts.iallocator)) |
248 |
elif self.opts.disk_template not in constants.DTS_NET_MIRROR: |
249 |
snode = None |
250 |
Log("- Add instance %s on node %s" % (instance, pnode)) |
251 |
else: |
252 |
Log("- Add instance %s on nodes %s/%s" % (instance, pnode, snode)) |
253 |
|
254 |
op = opcodes.OpCreateInstance(instance_name=instance, |
255 |
disk_size=self.opts.os_size, |
256 |
swap_size=self.opts.swap_size, |
257 |
disk_template=self.opts.disk_template, |
258 |
mode=constants.INSTANCE_CREATE, |
259 |
os_type=self.opts.os, |
260 |
pnode=pnode, |
261 |
snode=snode, |
262 |
start=True, |
263 |
ip_check=True, |
264 |
wait_for_sync=True, |
265 |
mac="auto", |
266 |
file_driver="loop", |
267 |
file_storage_dir=None, |
268 |
iallocator=self.opts.iallocator, |
269 |
beparams=self.bep, |
270 |
hvparams=self.hvp, |
271 |
) |
272 |
|
273 |
if self.opts.parallel: |
274 |
jobset.append([op]) |
275 |
# FIXME: here we should not append to to_rem uncoditionally, |
276 |
# but only when the job is successful |
277 |
self.to_rem.append(instance) |
278 |
else: |
279 |
self.ExecOp(op) |
280 |
self.to_rem.append(instance) |
281 |
if self.opts.parallel: |
282 |
self.ExecJobSet(jobset) |
283 |
|
284 |
def GrowDisks(self): |
285 |
"""Grow both the os and the swap disks by the requested amount, if any.""" |
286 |
for instance in self.instances: |
287 |
for disk in ['sda', 'sdb']: |
288 |
growth = getattr(self.opts, '%s_growth' % disk) |
289 |
if growth > 0: |
290 |
op = opcodes.OpGrowDisk(instance_name=instance, disk=disk, |
291 |
amount=growth, wait_for_sync=True) |
292 |
Log("- Increase %s's %s disk by %s MB" % (instance, disk, growth)) |
293 |
self.ExecOp(op) |
294 |
|
295 |
def ReplaceDisks1D8(self): |
296 |
"""Replace disks on primary and secondary for drbd8.""" |
297 |
for instance in self.instances: |
298 |
for mode in constants.REPLACE_DISK_SEC, constants.REPLACE_DISK_PRI: |
299 |
op = opcodes.OpReplaceDisks(instance_name=instance, |
300 |
mode=mode, |
301 |
disks=["sda", "sdb"]) |
302 |
Log("- Replace disks (%s) for instance %s" % (mode, instance)) |
303 |
self.ExecOp(op) |
304 |
|
305 |
def ReplaceDisks2(self): |
306 |
"""Replace secondary node.""" |
307 |
mode = constants.REPLACE_DISK_SEC |
308 |
|
309 |
mytor = izip(islice(cycle(self.nodes), 2, None), |
310 |
self.instances) |
311 |
for tnode, instance in mytor: |
312 |
if self.opts.iallocator: |
313 |
tnode = None |
314 |
op = opcodes.OpReplaceDisks(instance_name=instance, |
315 |
mode=mode, |
316 |
remote_node=tnode, |
317 |
iallocator=self.opts.iallocator, |
318 |
disks=["sda", "sdb"]) |
319 |
Log("- Replace secondary (%s) for instance %s" % (mode, instance)) |
320 |
self.ExecOp(op) |
321 |
|
322 |
def Failover(self): |
323 |
"""Failover the instances.""" |
324 |
|
325 |
for instance in self.instances: |
326 |
op = opcodes.OpFailoverInstance(instance_name=instance, |
327 |
ignore_consistency=False) |
328 |
|
329 |
Log("- Failover instance %s" % (instance)) |
330 |
self.ExecOp(op) |
331 |
|
332 |
def ImportExport(self): |
333 |
"""Export the instance, delete it, and import it back. |
334 |
|
335 |
""" |
336 |
|
337 |
mytor = izip(cycle(self.nodes), |
338 |
islice(cycle(self.nodes), 1, None), |
339 |
islice(cycle(self.nodes), 2, None), |
340 |
self.instances) |
341 |
|
342 |
for pnode, snode, enode, instance in mytor: |
343 |
|
344 |
if self.opts.iallocator: |
345 |
pnode = snode = None |
346 |
import_log_msg = ("- Import instance %s from node %s (iallocator: %s)" % |
347 |
(instance, enode, self.opts.iallocator)) |
348 |
elif self.opts.disk_template not in constants.DTS_NET_MIRROR: |
349 |
snode = None |
350 |
import_log_msg = ("- Import instance %s from node %s to node %s" % |
351 |
(instance, enode, pnode)) |
352 |
else: |
353 |
import_log_msg = ("- Import instance %s from node %s to nodes %s/%s" % |
354 |
(instance, enode, pnode, snode)) |
355 |
|
356 |
exp_op = opcodes.OpExportInstance(instance_name=instance, |
357 |
target_node=enode, |
358 |
shutdown=True) |
359 |
rem_op = opcodes.OpRemoveInstance(instance_name=instance, |
360 |
ignore_failures=True) |
361 |
nam_op = opcodes.OpQueryInstances(output_fields=["name"], |
362 |
names=[instance]) |
363 |
full_name = self.ExecOp(nam_op)[0][0] |
364 |
imp_dir = os.path.join(constants.EXPORT_DIR, full_name) |
365 |
imp_op = opcodes.OpCreateInstance(instance_name=instance, |
366 |
disk_size=self.opts.os_size, |
367 |
swap_size=self.opts.swap_size, |
368 |
disk_template=self.opts.disk_template, |
369 |
mode=constants.INSTANCE_IMPORT, |
370 |
src_node=enode, |
371 |
src_path=imp_dir, |
372 |
pnode=pnode, |
373 |
snode=snode, |
374 |
start=True, |
375 |
ip_check=True, |
376 |
wait_for_sync=True, |
377 |
mac="auto", |
378 |
file_storage_dir=None, |
379 |
file_driver=None, |
380 |
iallocator=self.opts.iallocator, |
381 |
beparams=self.bep, |
382 |
hvparams=self.hvp, |
383 |
) |
384 |
|
385 |
erem_op = opcodes.OpRemoveExport(instance_name=instance) |
386 |
|
387 |
Log("- Export instance %s to node %s" % (instance, enode)) |
388 |
self.ExecOp(exp_op) |
389 |
Log("- Remove instance %s" % (instance)) |
390 |
self.ExecOp(rem_op) |
391 |
self.to_rem.remove(instance) |
392 |
Log(import_log_msg) |
393 |
self.ExecOp(imp_op) |
394 |
Log("- Remove export of instance %s" % (instance)) |
395 |
self.ExecOp(erem_op) |
396 |
|
397 |
self.to_rem.append(instance) |
398 |
|
399 |
def StopInstance(self, instance): |
400 |
"""Stop given instance.""" |
401 |
op = opcodes.OpShutdownInstance(instance_name=instance) |
402 |
Log("- Shutdown instance %s" % instance) |
403 |
self.ExecOp(op) |
404 |
|
405 |
def StartInstance(self, instance): |
406 |
"""Start given instance.""" |
407 |
op = opcodes.OpStartupInstance(instance_name=instance, force=False) |
408 |
Log("- Start instance %s" % instance) |
409 |
self.ExecOp(op) |
410 |
|
411 |
def RenameInstance(self, instance, instance_new): |
412 |
"""Rename instance.""" |
413 |
op = opcodes.OpRenameInstance(instance_name=instance, |
414 |
new_name=instance_new) |
415 |
Log("- Rename instance %s to %s" % (instance, instance_new)) |
416 |
self.ExecOp(op) |
417 |
|
418 |
def StopStart(self): |
419 |
"""Stop/start the instances.""" |
420 |
for instance in self.instances: |
421 |
self.StopInstance(instance) |
422 |
self.StartInstance(instance) |
423 |
|
424 |
def Remove(self): |
425 |
"""Remove the instances.""" |
426 |
for instance in self.to_rem: |
427 |
op = opcodes.OpRemoveInstance(instance_name=instance, |
428 |
ignore_failures=True) |
429 |
Log("- Remove instance %s" % instance) |
430 |
self.ExecOp(op) |
431 |
|
432 |
|
433 |
def Rename(self): |
434 |
"""Rename the instances.""" |
435 |
rename = self.opts.rename |
436 |
for instance in self.instances: |
437 |
self.StopInstance(instance) |
438 |
self.RenameInstance(instance, rename) |
439 |
self.StartInstance(rename) |
440 |
self.StopInstance(rename) |
441 |
self.RenameInstance(rename, instance) |
442 |
self.StartInstance(instance) |
443 |
|
444 |
def BurninCluster(self): |
445 |
"""Test a cluster intensively. |
446 |
|
447 |
This will create instances and then start/stop/failover them. |
448 |
It is safe for existing instances but could impact performance. |
449 |
|
450 |
""" |
451 |
|
452 |
opts = self.opts |
453 |
|
454 |
Log("- Testing global parameters") |
455 |
|
456 |
if (len(self.nodes) == 1 and |
457 |
opts.disk_template not in (constants.DT_DISKLESS, constants.DT_PLAIN, |
458 |
constants.DT_FILE)): |
459 |
Log("When one node is available/selected the disk template must" |
460 |
" be 'diskless', 'file' or 'plain'") |
461 |
sys.exit(1) |
462 |
|
463 |
has_err = True |
464 |
try: |
465 |
self.CreateInstances() |
466 |
if opts.do_replace1 and opts.disk_template in constants.DTS_NET_MIRROR: |
467 |
self.ReplaceDisks1D8() |
468 |
if (opts.do_replace2 and len(self.nodes) > 2 and |
469 |
opts.disk_template in constants.DTS_NET_MIRROR) : |
470 |
self.ReplaceDisks2() |
471 |
|
472 |
if opts.disk_template != constants.DT_DISKLESS: |
473 |
self.GrowDisks() |
474 |
|
475 |
if opts.do_failover and opts.disk_template in constants.DTS_NET_MIRROR: |
476 |
self.Failover() |
477 |
|
478 |
if opts.do_importexport: |
479 |
self.ImportExport() |
480 |
|
481 |
if opts.do_startstop: |
482 |
self.StopStart() |
483 |
|
484 |
if opts.rename: |
485 |
self.Rename() |
486 |
|
487 |
has_err = False |
488 |
finally: |
489 |
if has_err: |
490 |
Log("Error detected: opcode buffer follows:\n\n") |
491 |
Log(self.GetFeedbackBuf()) |
492 |
Log("\n\n") |
493 |
self.Remove() |
494 |
|
495 |
return 0 |
496 |
|
497 |
|
498 |
def main(): |
499 |
"""Main function""" |
500 |
|
501 |
burner = Burner() |
502 |
return burner.BurninCluster() |
503 |
|
504 |
|
505 |
if __name__ == "__main__": |
506 |
main() |