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
|
import logging
import json
import re
import socket
from typing import cast, Dict, List, Any, Union, Optional, TYPE_CHECKING, Tuple
from mgr_module import NFS_POOL_NAME as POOL_NAME
from ceph.deployment.service_spec import NFSServiceSpec, PlacementSpec, IngressSpec
import orchestrator
from .exception import NFSInvalidOperation, ClusterNotFound
from .utils import (available_clusters, restart_nfs_service, conf_obj_name,
user_conf_obj_name)
from .export import NFSRados, exception_handler
if TYPE_CHECKING:
from nfs.module import Module
from mgr_module import MgrModule
log = logging.getLogger(__name__)
def resolve_ip(hostname: str) -> str:
try:
r = socket.getaddrinfo(hostname, None, flags=socket.AI_CANONNAME,
type=socket.SOCK_STREAM)
# pick first v4 IP, if present
for a in r:
if a[0] == socket.AF_INET:
return a[4][0]
return r[0][4][0]
except socket.gaierror as e:
raise NFSInvalidOperation(f"Cannot resolve IP for host {hostname}: {e}")
def create_ganesha_pool(mgr: 'MgrModule') -> None:
pool_list = [p['pool_name'] for p in mgr.get_osdmap().dump().get('pools', [])]
if POOL_NAME not in pool_list:
mgr.check_mon_command({'prefix': 'osd pool create', 'pool': POOL_NAME})
mgr.check_mon_command({'prefix': 'osd pool application enable',
'pool': POOL_NAME,
'app': 'nfs'})
log.debug("Successfully created nfs-ganesha pool %s", POOL_NAME)
class NFSCluster:
def __init__(self, mgr: 'Module') -> None:
self.mgr = mgr
def _call_orch_apply_nfs(
self,
cluster_id: str,
placement: Optional[str],
virtual_ip: Optional[str] = None,
port: Optional[int] = None,
) -> None:
if not port:
port = 2049 # default nfs port
if virtual_ip:
# nfs + ingress
# run NFS on non-standard port
spec = NFSServiceSpec(service_type='nfs', service_id=cluster_id,
placement=PlacementSpec.from_string(placement),
# use non-default port so we don't conflict with ingress
port=10000 + port) # semi-arbitrary, fix me someday
completion = self.mgr.apply_nfs(spec)
orchestrator.raise_if_exception(completion)
ispec = IngressSpec(service_type='ingress',
service_id='nfs.' + cluster_id,
backend_service='nfs.' + cluster_id,
frontend_port=port,
monitor_port=7000 + port, # semi-arbitrary, fix me someday
virtual_ip=virtual_ip)
completion = self.mgr.apply_ingress(ispec)
orchestrator.raise_if_exception(completion)
else:
# standalone nfs
spec = NFSServiceSpec(service_type='nfs', service_id=cluster_id,
placement=PlacementSpec.from_string(placement),
port=port)
completion = self.mgr.apply_nfs(spec)
orchestrator.raise_if_exception(completion)
log.debug("Successfully deployed nfs daemons with cluster id %s and placement %s",
cluster_id, placement)
def create_empty_rados_obj(self, cluster_id: str) -> None:
common_conf = conf_obj_name(cluster_id)
self._rados(cluster_id).write_obj('', conf_obj_name(cluster_id))
log.info("Created empty object:%s", common_conf)
def delete_config_obj(self, cluster_id: str) -> None:
self._rados(cluster_id).remove_all_obj()
log.info("Deleted %s object and all objects in %s",
conf_obj_name(cluster_id), cluster_id)
def create_nfs_cluster(
self,
cluster_id: str,
placement: Optional[str],
virtual_ip: Optional[str],
ingress: Optional[bool] = None,
port: Optional[int] = None,
) -> Tuple[int, str, str]:
try:
if virtual_ip and not ingress:
raise NFSInvalidOperation('virtual_ip can only be provided with ingress enabled')
if not virtual_ip and ingress:
raise NFSInvalidOperation('ingress currently requires a virtual_ip')
invalid_str = re.search('[^A-Za-z0-9-_.]', cluster_id)
if invalid_str:
raise NFSInvalidOperation(f"cluster id {cluster_id} is invalid. "
f"{invalid_str.group()} is char not permitted")
create_ganesha_pool(self.mgr)
self.create_empty_rados_obj(cluster_id)
if cluster_id not in available_clusters(self.mgr):
self._call_orch_apply_nfs(cluster_id, placement, virtual_ip, port)
return 0, "NFS Cluster Created Successfully", ""
return 0, "", f"{cluster_id} cluster already exists"
except Exception as e:
return exception_handler(e, f"NFS Cluster {cluster_id} could not be created")
def delete_nfs_cluster(self, cluster_id: str) -> Tuple[int, str, str]:
try:
cluster_list = available_clusters(self.mgr)
if cluster_id in cluster_list:
self.mgr.export_mgr.delete_all_exports(cluster_id)
completion = self.mgr.remove_service('ingress.nfs.' + cluster_id)
orchestrator.raise_if_exception(completion)
completion = self.mgr.remove_service('nfs.' + cluster_id)
orchestrator.raise_if_exception(completion)
self.delete_config_obj(cluster_id)
return 0, "NFS Cluster Deleted Successfully", ""
return 0, "", "Cluster does not exist"
except Exception as e:
return exception_handler(e, f"Failed to delete NFS Cluster {cluster_id}")
def list_nfs_cluster(self) -> Tuple[int, str, str]:
try:
return 0, '\n'.join(available_clusters(self.mgr)), ""
except Exception as e:
return exception_handler(e, "Failed to list NFS Cluster")
def _show_nfs_cluster_info(self, cluster_id: str) -> Dict[str, Any]:
completion = self.mgr.list_daemons(daemon_type='nfs')
# Here completion.result is a list DaemonDescription objects
clusters = orchestrator.raise_if_exception(completion)
backends: List[Dict[str, Union[Any]]] = []
for cluster in clusters:
if cluster_id == cluster.service_id():
assert cluster.hostname
try:
if cluster.ip:
ip = cluster.ip
else:
c = self.mgr.get_hosts()
orchestrator.raise_if_exception(c)
hosts = [h for h in c.result or []
if h.hostname == cluster.hostname]
if hosts:
ip = resolve_ip(hosts[0].addr)
else:
# sigh
ip = resolve_ip(cluster.hostname)
backends.append({
"hostname": cluster.hostname,
"ip": ip,
"port": cluster.ports[0] if cluster.ports else None
})
except orchestrator.OrchestratorError:
continue
r: Dict[str, Any] = {
'virtual_ip': None,
'backend': backends,
}
sc = self.mgr.describe_service(service_type='ingress')
services = orchestrator.raise_if_exception(sc)
for i in services:
spec = cast(IngressSpec, i.spec)
if spec.backend_service == f'nfs.{cluster_id}':
r['virtual_ip'] = i.virtual_ip.split('/')[0] if i.virtual_ip else None
if i.ports:
r['port'] = i.ports[0]
if len(i.ports) > 1:
r['monitor_port'] = i.ports[1]
log.debug("Successfully fetched %s info: %s", cluster_id, r)
return r
def show_nfs_cluster_info(self, cluster_id: Optional[str] = None) -> Tuple[int, str, str]:
try:
info_res = {}
if cluster_id:
cluster_ls = [cluster_id]
else:
cluster_ls = available_clusters(self.mgr)
for cluster_id in cluster_ls:
res = self._show_nfs_cluster_info(cluster_id)
if res:
info_res[cluster_id] = res
return (0, json.dumps(info_res, indent=4), '')
except Exception as e:
return exception_handler(e, "Failed to show info for cluster")
def get_nfs_cluster_config(self, cluster_id: str) -> Tuple[int, str, str]:
try:
if cluster_id in available_clusters(self.mgr):
rados_obj = self._rados(cluster_id)
conf = rados_obj.read_obj(user_conf_obj_name(cluster_id))
return 0, conf or "", ""
raise ClusterNotFound()
except Exception as e:
return exception_handler(e, f"Fetching NFS-Ganesha Config failed for {cluster_id}")
def set_nfs_cluster_config(self, cluster_id: str, nfs_config: str) -> Tuple[int, str, str]:
try:
if cluster_id in available_clusters(self.mgr):
rados_obj = self._rados(cluster_id)
if rados_obj.check_user_config():
return 0, "", "NFS-Ganesha User Config already exists"
rados_obj.write_obj(nfs_config, user_conf_obj_name(cluster_id),
conf_obj_name(cluster_id))
log.debug("Successfully saved %s's user config: \n %s", cluster_id, nfs_config)
restart_nfs_service(self.mgr, cluster_id)
return 0, "NFS-Ganesha Config Set Successfully", ""
raise ClusterNotFound()
except NotImplementedError:
return 0, "NFS-Ganesha Config Added Successfully "\
"(Manual Restart of NFS PODS required)", ""
except Exception as e:
return exception_handler(e, f"Setting NFS-Ganesha Config failed for {cluster_id}")
def reset_nfs_cluster_config(self, cluster_id: str) -> Tuple[int, str, str]:
try:
if cluster_id in available_clusters(self.mgr):
rados_obj = self._rados(cluster_id)
if not rados_obj.check_user_config():
return 0, "", "NFS-Ganesha User Config does not exist"
rados_obj.remove_obj(user_conf_obj_name(cluster_id),
conf_obj_name(cluster_id))
restart_nfs_service(self.mgr, cluster_id)
return 0, "NFS-Ganesha Config Reset Successfully", ""
raise ClusterNotFound()
except NotImplementedError:
return 0, "NFS-Ganesha Config Removed Successfully "\
"(Manual Restart of NFS PODS required)", ""
except Exception as e:
return exception_handler(e, f"Resetting NFS-Ganesha Config failed for {cluster_id}")
def _rados(self, cluster_id: str) -> NFSRados:
"""Return a new NFSRados object for the given cluster id."""
return NFSRados(self.mgr.rados, cluster_id)
|
