summaryrefslogtreecommitdiffstats
path: root/qa/tasks/rgw.py
blob: e747426c374939f93ed9bbb37b21f081cbb77aee (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
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
"""
rgw routines
"""
import argparse
import contextlib
import logging

from teuthology.orchestra import run
from teuthology import misc as teuthology
from teuthology import contextutil
from teuthology.exceptions import ConfigError
from tasks.util import get_remote_for_role
from tasks.util.rgw import rgwadmin, wait_for_radosgw
from tasks.util.rados import (create_ec_pool,
                              create_replicated_pool,
                              create_cache_pool)

log = logging.getLogger(__name__)

class RGWEndpoint:
    def __init__(self, hostname=None, port=None, cert=None, dns_name=None, website_dns_name=None):
        self.hostname = hostname
        self.port = port
        self.cert = cert
        self.dns_name = dns_name
        self.website_dns_name = website_dns_name

    def url(self):
        proto = 'https' if self.cert else 'http'
        return '{proto}://{hostname}:{port}/'.format(proto=proto, hostname=self.hostname, port=self.port)

@contextlib.contextmanager
def start_rgw(ctx, config, clients):
    """
    Start rgw on remote sites.
    """
    log.info('Starting rgw...')
    testdir = teuthology.get_testdir(ctx)
    for client in clients:
        (remote,) = ctx.cluster.only(client).remotes.keys()
        cluster_name, daemon_type, client_id = teuthology.split_role(client)
        client_with_id = daemon_type + '.' + client_id
        client_with_cluster = cluster_name + '.' + client_with_id

        client_config = config.get(client)
        if client_config is None:
            client_config = {}
        log.info("rgw %s config is %s", client, client_config)
        cmd_prefix = [
            'sudo',
            'adjust-ulimits',
            'ceph-coverage',
            '{tdir}/archive/coverage'.format(tdir=testdir),
            'daemon-helper',
            'term',
            ]

        rgw_cmd = ['radosgw']

        log.info("Using %s as radosgw frontend", ctx.rgw.frontend)

        endpoint = ctx.rgw.role_endpoints[client]
        frontends = ctx.rgw.frontend
        frontend_prefix = client_config.get('frontend_prefix', None)
        if frontend_prefix:
            frontends += ' prefix={pfx}'.format(pfx=frontend_prefix)

        if endpoint.cert:
            # add the ssl certificate path
            frontends += ' ssl_certificate={}'.format(endpoint.cert.certificate)
            if ctx.rgw.frontend == 'civetweb':
                frontends += ' port={}s'.format(endpoint.port)
            else:
                frontends += ' ssl_port={}'.format(endpoint.port)
        else:
            frontends += ' port={}'.format(endpoint.port)

        rgw_cmd.extend([
            '--rgw-frontends', frontends,
            '-n', client_with_id,
            '--cluster', cluster_name,
            '-k', '/etc/ceph/{client_with_cluster}.keyring'.format(client_with_cluster=client_with_cluster),
            '--log-file',
            '/var/log/ceph/rgw.{client_with_cluster}.log'.format(client_with_cluster=client_with_cluster),
            '--rgw_ops_log_socket_path',
            '{tdir}/rgw.opslog.{client_with_cluster}.sock'.format(tdir=testdir,
                                                     client_with_cluster=client_with_cluster)
	    ])

        keystone_role = client_config.get('use-keystone-role', None)
        if keystone_role is not None:
            if not ctx.keystone:
                raise ConfigError('rgw must run after the keystone task')
            url = 'http://{host}:{port}/v1/KEY_$(tenant_id)s'.format(host=endpoint.hostname,
                                                                     port=endpoint.port)
            ctx.keystone.create_endpoint(ctx, keystone_role, 'swift', url)

            keystone_host, keystone_port = \
                ctx.keystone.public_endpoints[keystone_role]
            rgw_cmd.extend([
                '--rgw_keystone_url',
                'http://{khost}:{kport}'.format(khost=keystone_host,
                                                kport=keystone_port),
                ])


        if client_config.get('dns-name') is not None:
            rgw_cmd.extend(['--rgw-dns-name', endpoint.dns_name])
        if client_config.get('dns-s3website-name') is not None:
            rgw_cmd.extend(['--rgw-dns-s3website-name', endpoint.website_dns_name])

        rgw_cmd.extend([
            '--foreground',
            run.Raw('|'),
            'sudo',
            'tee',
            '/var/log/ceph/rgw.{client_with_cluster}.stdout'.format(tdir=testdir,
                                                       client_with_cluster=client_with_cluster),
            run.Raw('2>&1'),
            ])

        if client_config.get('valgrind'):
            cmd_prefix = teuthology.get_valgrind_args(
                testdir,
                client_with_cluster,
                cmd_prefix,
                client_config.get('valgrind')
                )

        run_cmd = list(cmd_prefix)
        run_cmd.extend(rgw_cmd)

        ctx.daemons.add_daemon(
            remote, 'rgw', client_with_id,
            cluster=cluster_name,
            args=run_cmd,
            logger=log.getChild(client),
            stdin=run.PIPE,
            wait=False,
            )

    # XXX: add_daemon() doesn't let us wait until radosgw finishes startup
    for client in clients:
        endpoint = ctx.rgw.role_endpoints[client]
        url = endpoint.url()
        log.info('Polling {client} until it starts accepting connections on {url}'.format(client=client, url=url))
        (remote,) = ctx.cluster.only(client).remotes.keys()
        wait_for_radosgw(url, remote)

    try:
        yield
    finally:
        for client in clients:
            cluster_name, daemon_type, client_id = teuthology.split_role(client)
            client_with_id = daemon_type + '.' + client_id
            client_with_cluster = cluster_name + '.' + client_with_id
            ctx.daemons.get_daemon('rgw', client_with_id, cluster_name).stop()
            ctx.cluster.only(client).run(
                args=[
                    'rm',
                    '-f',
                    '{tdir}/rgw.opslog.{client}.sock'.format(tdir=testdir,
                                                             client=client_with_cluster),
                    ],
                )

def assign_endpoints(ctx, config, default_cert):
    role_endpoints = {}
    for role, client_config in config.items():
        client_config = client_config or {}
        remote = get_remote_for_role(ctx, role)

        cert = client_config.get('ssl certificate', default_cert)
        if cert:
            # find the certificate created by the ssl task
            if not hasattr(ctx, 'ssl_certificates'):
                raise ConfigError('rgw: no ssl task found for option "ssl certificate"')
            ssl_certificate = ctx.ssl_certificates.get(cert, None)
            if not ssl_certificate:
                raise ConfigError('rgw: missing ssl certificate "{}"'.format(cert))
        else:
            ssl_certificate = None

        port = client_config.get('port', 443 if ssl_certificate else 80)

        # if dns-name is given, use it as the hostname (or as a prefix)
        dns_name = client_config.get('dns-name', '')
        if len(dns_name) == 0 or dns_name.endswith('.'):
            dns_name += remote.hostname

        website_dns_name = client_config.get('dns-s3website-name')
        if website_dns_name is not None and (len(website_dns_name) == 0 or website_dns_name.endswith('.')):
            website_dns_name += remote.hostname

        role_endpoints[role] = RGWEndpoint(remote.hostname, port, ssl_certificate, dns_name, website_dns_name)

    return role_endpoints

@contextlib.contextmanager
def create_pools(ctx, clients):
    """Create replicated or erasure coded data pools for rgw."""

    log.info('Creating data pools')
    for client in clients:
        log.debug("Obtaining remote for client {}".format(client))
        (remote,) = ctx.cluster.only(client).remotes.keys()
        data_pool = 'default.rgw.buckets.data'
        cluster_name, daemon_type, client_id = teuthology.split_role(client)

        if ctx.rgw.ec_data_pool:
            create_ec_pool(remote, data_pool, client, ctx.rgw.data_pool_pg_size,
                           ctx.rgw.erasure_code_profile, cluster_name, 'rgw')
        else:
            create_replicated_pool(remote, data_pool, ctx.rgw.data_pool_pg_size, cluster_name, 'rgw')

        index_pool = 'default.rgw.buckets.index'
        create_replicated_pool(remote, index_pool, ctx.rgw.index_pool_pg_size, cluster_name, 'rgw')

        if ctx.rgw.cache_pools:
            create_cache_pool(remote, data_pool, data_pool + '.cache', 64,
                              64*1024*1024, cluster_name)
    log.debug('Pools created')
    yield

@contextlib.contextmanager
def configure_compression(ctx, clients, compression):
    """ set a compression type in the default zone placement """
    log.info('Configuring compression type = %s', compression)
    for client in clients:
        # XXX: the 'default' zone and zonegroup aren't created until we run RGWRados::init_complete().
        # issue a 'radosgw-admin user list' command to trigger this
        rgwadmin(ctx, client, cmd=['user', 'list'], check_status=True)

        rgwadmin(ctx, client,
                cmd=['zone', 'placement', 'modify', '--rgw-zone', 'default',
                     '--placement-id', 'default-placement',
                     '--compression', compression],
                check_status=True)
    yield

@contextlib.contextmanager
def configure_storage_classes(ctx, clients, storage_classes):
    """ set a compression type in the default zone placement """

    sc = [s.strip() for s in storage_classes.split(',')]

    for client in clients:
        # XXX: the 'default' zone and zonegroup aren't created until we run RGWRados::init_complete().
        # issue a 'radosgw-admin user list' command to trigger this
        rgwadmin(ctx, client, cmd=['user', 'list'], check_status=True)

        for storage_class in sc:
            log.info('Configuring storage class type = %s', storage_class)
            rgwadmin(ctx, client,
                    cmd=['zonegroup', 'placement', 'add',
                        '--rgw-zone', 'default',
                        '--placement-id', 'default-placement',
                        '--storage-class', storage_class],
                    check_status=True)
            rgwadmin(ctx, client,
                    cmd=['zone', 'placement', 'add',
                        '--rgw-zone', 'default',
                        '--placement-id', 'default-placement',
                        '--storage-class', storage_class,
                        '--data-pool', 'default.rgw.buckets.data.' + storage_class.lower()],
                    check_status=True)
    yield

@contextlib.contextmanager
def task(ctx, config):
    """
    For example, to run rgw on all clients::

        tasks:
        - ceph:
        - rgw:

    To only run on certain clients::

        tasks:
        - ceph:
        - rgw: [client.0, client.3]

    or

        tasks:
        - ceph:
        - rgw:
            client.0:
            client.3:

    To run radosgw through valgrind:

        tasks:
        - ceph:
        - rgw:
            client.0:
              valgrind: [--tool=memcheck]
            client.3:
              valgrind: [--tool=memcheck]

    To configure data or index pool pg_size:

        overrides:
          rgw:
            data_pool_pg_size: 256
            index_pool_pg_size: 128
    """
    if config is None:
        config = dict(('client.{id}'.format(id=id_), None)
                      for id_ in teuthology.all_roles_of_type(
                          ctx.cluster, 'client'))
    elif isinstance(config, list):
        config = dict((name, None) for name in config)

    clients = config.keys() # http://tracker.ceph.com/issues/20417

    overrides = ctx.config.get('overrides', {})
    teuthology.deep_merge(config, overrides.get('rgw', {}))

    ctx.rgw = argparse.Namespace()

    ctx.rgw.ec_data_pool = bool(config.pop('ec-data-pool', False))
    ctx.rgw.erasure_code_profile = config.pop('erasure_code_profile', {})
    ctx.rgw.cache_pools = bool(config.pop('cache-pools', False))
    ctx.rgw.frontend = config.pop('frontend', 'civetweb')
    ctx.rgw.compression_type = config.pop('compression type', None)
    ctx.rgw.storage_classes = config.pop('storage classes', None)
    default_cert = config.pop('ssl certificate', None)
    ctx.rgw.data_pool_pg_size = config.pop('data_pool_pg_size', 64)
    ctx.rgw.index_pool_pg_size = config.pop('index_pool_pg_size', 64)
    ctx.rgw.config = config

    log.debug("config is {}".format(config))
    log.debug("client list is {}".format(clients))

    ctx.rgw.role_endpoints = assign_endpoints(ctx, config, default_cert)

    subtasks = [
        lambda: create_pools(ctx=ctx, clients=clients),
    ]
    if ctx.rgw.compression_type:
        subtasks.extend([
            lambda: configure_compression(ctx=ctx, clients=clients,
                                          compression=ctx.rgw.compression_type),
        ])
    if ctx.rgw.storage_classes:
        subtasks.extend([
            lambda: configure_storage_classes(ctx=ctx, clients=clients,
                                              storage_classes=ctx.rgw.storage_classes),
        ])
    subtasks.extend([
        lambda: start_rgw(ctx=ctx, config=config, clients=clients),
    ])

    with contextutil.nested(*subtasks):
        yield