1MOUNT.CEPH(8) Ceph MOUNT.CEPH(8)
2
3
4
6 mount.ceph - mount a Ceph file system
7
9 mount.ceph [mon1_socket,mon2_socket,...]:/[subdir] dir [
10 -o options ]
11
12
14 mount.ceph is a helper for mounting the Ceph file system on a Linux
15 host. It serves to resolve monitor hostname(s) into IP addresses and
16 read authentication keys from disk; the Linux kernel client component
17 does most of the real work. In fact, it is possible to mount a non-au‐
18 thenticated Ceph file system without mount.ceph by specifying monitor
19 address(es) by IP:
20
21 mount -t ceph 1.2.3.4:/ /mnt/mycephfs
22
23 The first argument is the device part of the mount command. It includes
24 host's socket and path within CephFS that will be mounted at the mount
25 point. The socket, obviously, takes the form ip_address[:port]. If the
26 port is not specified, the Ceph default of 6789 is assumed. Multiple
27 monitor addresses can be passed by separating them by commas. Only one
28 monitor is needed to mount successfully; the client will learn about
29 all monitors from any responsive monitor. However, it is a good idea to
30 specify more than one in case the one happens to be down at the time of
31 mount.
32
33 If the host portion of the device is left blank, then mount.ceph will
34 attempt to determine monitor addresses using local configuration files
35 and/or DNS SRV records. In similar way, if authentication is enabled on
36 Ceph cluster (which is done using CephX) and options secret and secret‐
37 file are not specified in the command, the mount helper will spawn a
38 child process that will use the standard Ceph library routines to find
39 a keyring and fetch the secret from it.
40
41 A sub-directory of the file system can be mounted by specifying the
42 (absolute) path to the sub-directory right after ":" after the socket
43 in the device part of the mount command.
44
45 Mount helper application conventions dictate that the first two options
46 are device to be mounted and the mountpoint for that device. Options
47 must be passed only after these fixed arguments.
48
50 Basic
51 conf Path to a ceph.conf file. This is used to initialize the Ceph
52 context for autodiscovery of monitor addresses and auth secrets.
53 The default is to use the standard search path for ceph.conf
54 files.
55
56 command
57 fs=<fs-name> Specify the non-default file system to be mounted.
58 Not passing this option mounts the default file system.
59
60 command
61 mds_namespace=<fs-name> A synonym of "fs=" and its use is depre‐
62 cated.
63
64 mount_timeout
65 int (seconds), Default: 60
66
67 ms_mode=<legacy|crc|secure|prefer-crc|prefer-secure>
68 Set the connection mode that the client uses for transport. The
69 available modes are:
70
71 • legacy: use messenger v1 protocol to talk to the cluster
72
73 • crc: use messenger v2, without on-the-wire encryption
74
75 • secure: use messenger v2, with on-the-wire encryption
76
77 • prefer-crc: crc mode, if denied agree to secure mode
78
79 • prefer-secure: secure mode, if denied agree to crc mode
80
81 name RADOS user to authenticate as when using CephX. Default: guest
82
83 secret secret key for use with CephX. This option is insecure because
84 it exposes the secret on the command line. To avoid this, use
85 the secretfile option.
86
87 secretfile
88 path to file containing the secret key to use with CephX
89
90 recover_session=<no|clean>
91 Set auto reconnect mode in the case where the client is block‐
92 listed. The available modes are no and clean. The default is no.
93
94 •
95
96 no: never attempt to reconnect when client detects that it has
97 been
98 blocklisted. Blocklisted clients will not attempt to
99 reconnect and their operations will fail too.
100
101 • clean: client reconnects to the Ceph cluster automatically
102 when it detects that it has been blocklisted. During recon‐
103 nect, client drops dirty data/metadata, invalidates page
104 caches and writable file handles. After reconnect, file locks
105 become stale because the MDS loses track of them. If an inode
106 contains any stale file locks, read/write on the inode is not
107 allowed until applications release all stale file locks.
108
109 Advanced
110 cap_release_safety
111 int, Default: calculated
112
113 caps_wanted_delay_max
114 int, cap release delay, Default: 60
115
116 caps_wanted_delay_min
117 int, cap release delay, Default: 5
118
119 dirstat
120 funky cat dirname for stats, Default: off
121
122 nodirstat
123 no funky cat dirname for stats
124
125 ip my ip
126
127 noasyncreaddir
128 no dcache readdir
129
130 nocrc no data crc on writes
131
132 noshare
133 create a new client instance, instead of sharing an existing in‐
134 stance of a client mounting the same cluster
135
136 osdkeepalive
137 int, Default: 5
138
139 osd_idle_ttl
140 int (seconds), Default: 60
141
142 rasize int (bytes), max readahead. Default: 8388608 (8192*1024)
143
144 rbytes Report the recursive size of the directory contents for st_size
145 on directories. Default: off
146
147 norbytes
148 Do not report the recursive size of the directory contents for
149 st_size on directories.
150
151 readdir_max_bytes
152 int, Default: 524288 (512*1024)
153
154 readdir_max_entries
155 int, Default: 1024
156
157 rsize int (bytes), max read size. Default: 16777216 (16*1024*1024)
158
159 snapdirname
160 string, set the name of the hidden snapdir. Default: .snap
161
162 write_congestion_kb
163 int (kb), max writeback in flight. scale with available memory.
164 Default: calculated from available memory
165
166 wsize int (bytes), max write size. Default: 16777216 (16*1024*1024)
167 (writeback uses smaller of wsize and stripe unit)
168
169 wsync Execute all namespace operations synchronously. This ensures
170 that the namespace operation will only complete after receiving
171 a reply from the MDS. This is the default.
172
173 nowsync
174 Allow the client to do namespace operations asynchronously. When
175 this option is enabled, a namespace operation may complete be‐
176 fore the MDS replies, if it has sufficient capabilities to do
177 so.
178
180 Mount the full file system:
181
182 mount.ceph :/ /mnt/mycephfs
183
184 Assuming mount.ceph is installed properly, it should be automatically
185 invoked by mount(8):
186
187 mount -t ceph :/ /mnt/mycephfs
188
189 Mount only part of the namespace/file system:
190
191 mount.ceph :/some/directory/in/cephfs /mnt/mycephfs
192
193 Mount non-default FS, in case cluster has multiple FSs::
194 mount -t ceph :/ /mnt/mycephfs2 -o fs=mycephfs2
195
196 or
197
198 mount -t ceph :/ /mnt/mycephfs2 -o mds_namespace=mycephfs2 #
199 This option name is deprecated.
200
201 Pass the monitor host's IP address, optionally:
202
203 mount.ceph 192.168.0.1:/ /mnt/mycephfs
204
205 Pass the port along with IP address if it's running on a non-standard
206 port:
207
208 mount.ceph 192.168.0.1:7000:/ /mnt/mycephfs
209
210 If there are multiple monitors, passes addresses separated by a comma:
211
212 mount.ceph 192.168.0.1,192.168.0.2,192.168.0.3:/ /mnt/mycephfs
213
214 If authentication is enabled on Ceph cluster:
215
216 mount.ceph :/ /mnt/mycephfs -o name=fs_username
217
218 Pass secret key for CephX user optionally:
219
220 mount.ceph :/ /mnt/mycephfs -o name=fs_username,secret=AQATSKdNGBnwLhAAnNDKnH65FmVKpXZJVasUeQ==
221
222 Pass file containing secret key to avoid leaving secret key in shell's
223 command history:
224
225 mount.ceph :/ /mnt/mycephfs -o name=fs_username,secretfile=/etc/ceph/fs_username.secret
226
228 mount.ceph is part of Ceph, a massively scalable, open-source, distrib‐
229 uted storage system. Please refer to the Ceph documentation at
230 http://ceph.com/docs for more information.
231
233 The recover_session= option was added to mainline Linux kernels in
234 v5.4. wsync and nowsync were added in v5.7.
235
237 ceph-fuse(8), ceph(8)
238
240 2010-2022, Inktank Storage, Inc. and contributors. Licensed under Cre‐
241 ative Commons Attribution Share Alike 3.0 (CC-BY-SA-3.0)
242
243
244
245
246dev Jun 22, 2022 MOUNT.CEPH(8)