2 .\" Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
4 .\" Copyright (c) 2012, 2017, Intel Corporation.
6 .\" This file may be copied under the terms of the GNU Public License v2.
8 .TH mount.lustre 8 "2008 Mar 15" Lustre "configuration utilities"
10 mount.lustre \- start a Lustre client or target service
13 .BI "mount \-t lustre [\-o " options "] " "<block_device> <server_mountpoint>"
15 .BI "mount \-t lustre [\-o " options "] " "<mgsname>:/<fsname> <client_mountpoint>"
18 is used to start a Lustre client or target service. This program should not
19 be called directly; rather it is a helper program invoked through
23 filesystem type is used. Lustre clients and targets are stopped by using the
28 There are two forms for the device option passed to the
30 command, depending on whether a client or a target service is started:
32 .IR <mgsname> :/ <fsname>[/<subdir>]
33 mounts the Lustre filesystem named
35 (optionally starting at subdirectory
37 within the filesystem, if specified) on the client at the directory
38 .IR client_mountpoint ,
39 by contacting the Management Service at
43 is defined below. A client filesystem will typically be listed in
45 for automatic mount at boot time, is usable like any local filesystem, and
46 provides a full POSIX-compilant interface.
49 starts the target service defined by the
51 command on the physical disk or ZFS dataset given by
55 may be specified using
57 to find the first block device with that label (e.g.
61 option. Care should be taken if there is a device-level backup of
62 the target filesystem, which would have a duplicate label and UUID if it is
65 or similar. The mounted target service filesystem at
69 operations and appears in
71 to show the device is in use.
74 .BI <mgsname>:= <mgsnode>[:<mgsnode>]
77 may be a colon-separated list of
79 names where the MGS service may run. Multiple
81 values can be specified if the MGS service is configures for HA failover
82 and may be running on any one of the nodes.
84 .BI <mgsnode>:= <mgsnid>[,<mgsnid>]
87 may specify a comma-separated list of NIDs, if there are different
88 LNet interfaces for the same
92 Specifies the encryption flavour for the initial network RPC connection to
93 the MGS node. Non-security flavors are:
98 which respectively disable, or have no encryption or integrity features for
99 testing purposes. Kerberos flavors are:
105 Shared-secret key flavors are:
113 for more details. The security flavour for client-to-server connections is
114 specified in the filesystem configuration that the client fetches from the MGS.
116 .BI skpath= file|directory
117 Path to a file or directory with the keyfile(s) to load for this mount command.
118 Keys are inserted into the KEY_SPEC_SESSION_KEYRING keyring with a description
119 containing "lustre:" and a suffix which depends on whether the context of the
120 mount command is for an MGS, MDT/OST, or client.
121 This option is only available when built with --enable-gss.
124 Start a client or MDT with a (colon-separated) list of known inactive OSTs.
126 In addition to the standard options listed in
128 Lustre understands the following
133 Enable full distributed
135 support, coherent across all client nodes also using this mount option. This
136 is useful if applications need coherent userspace file locking across multiple
137 client nodes, but also imposes communications overhead in order to maintain
138 locking consistency between client nodes.
143 support, using only client-local file locking. This is faster than mounting
146 option, and can be used for applications that depend on functioning
148 but run only on a single node.
153 support entirely, and is the default option. Applications calling
157 error. It is up to the administrator to choose either the
161 mount option based on their requirements. It is possible to mount clients
162 with different options, and only those mounted with
164 will be coherent amongst each other.
173 to return even if some OST or MDT is unresponsive or has been temporarily
174 or permanently disabled in the configuration. This avoids blocking until
175 all of the targets are available. This is the default since Lustre 2.9.0.
180 block until all OSTs and MDTs are available and have returned space usage.
183 Enable get/set of extended attributes by regular users. See the
188 Disable use of extended attributes by regular users. Root and system processes can still use extended attributes.
191 Force a client to keep pinging even if servers have enabled suppress_pings.
194 Enable mount/remount/umount console messages.
197 Disable mount/remount/umount console messages.
200 Enable FID to path translation by regular users via
204 Disable FID to path translation by regular users. Root and process with
205 CAP_DAC_READ_SEARCH can still perform FID to path translation.
208 Limit connections from the client to be on the network NID specified by 'net'.
209 \'net\' designates a single network NID, like 'o2ib2' or 'tcp1'.
210 This option can be useful in case of several Lustre client mount
211 points on the same node, with each mount point using a different
212 network. It is also interesting when running Lustre clients from
213 containers, by restricting each container to a specific network.
216 Warning! 'network' option is incompatible with LNet Dynamic Peer Discovery.
217 If you want to restrict client NID, please make sure LNet Dynamic Peer Discovery
221 In addition to the standard mount options and backing disk type
222 (e.g. ldiskfs) options listed in
224 Lustre understands the following
229 Enable POSIX Access Control List support for all clients. See the
234 Only start the MGC (and MGS, if co-located) for a target service,
235 and not the actual MDS or OSS service.
238 Start an MDT with a co-located MGS without starting the MGS service.
241 Do not trigger OI scrub automatically when an inconsistency is detected.
242 It is still possible to start explicitly using the
243 .BR "lctl lfsck_start" .
246 Do not resume the former paused/crashed LFSCK automatically when mounting.
249 Abort client recovery and start the target service immediately.
251 .BI md_stripe_cache_size
252 Sets the stripe cache size for server side disk with a striped raid
256 Automatically Sets the block device parameter of 'max_sectors_kb' for the
257 MDT or OST target. When max_sectors_kb isn't specified, that parameter for
258 block device will be set to same as it's own 'max_hw_sectors_kb' (up to a
259 maximum of 16M), this is default behavior suited for most users. When
260 max_sectors_kb is specified as zero, the old parameter value will be kept.
261 When max_sectors_kb is specified as a positive number, the parameter will
262 be set to this number arbitrarily.
264 .BI recovery_time_soft= timeout
265 Allow 'timeout' seconds for clients to reconnect for recovery after a server
266 crash. This timeout will be incrementally extended if it is about to expire
267 and the server is still handling new connections from recoverable clients.
268 The default soft recovery timeout is set to 300 seconds (5 minutes).
270 .BI recovery_time_hard= timeout
271 The server will be allowed to incrementally extend its timeout up to a hard
272 maximum of 'timeout' seconds. The default hard recovery timeout is set to
273 900 seconds (15 minutes).
276 .B mount -t lustre cfs21@tcp0:/testfs /mnt/myfilesystem
277 Start a client for the Lustre filesystem
280 .BR /mnt/myfilesystem .
281 The Management Service is running on a node reachable via NID
284 .B mount -t lustre cfs21@tcp0:/testfs/dir /mnt/myfilesystem
285 Like above example, but mount subdirectory
289 .B mount -t lustre mgs1@tcp0,mgs1ib@o2ib0:mgs2@tcp0,mgs2ib@o2ib0:/testfs /mnt/fs
290 Like above example, but the Management Service is running on one of the service
294 .B mgs2, which are two different hosts separated by a colon and
295 served as a failover pair. Lustre tries the first one, and if that fails, it
296 tries the second one. On each service node, the comma-separated NIDs refer to
297 different interfaces on the same host, and the Lustre client chooses the best
298 one for communication based on which network interfaces are available locally.
300 .B mount -t lustre /dev/sda1 /mnt/test/mdt
301 Start the Lustre metadata target service from
306 .B mount -t lustre -L testfs-MDT0000 -o abort_recov /mnt/test/mdt
309 service (by using the disk label), but aborts the recovery process if
310 all of the clients are known to be unavailable.
312 Not very many mount options can be changed with
323 .BR tunefs.lustre (8),