1mount.gfs2(8) System Manager's Manual mount.gfs2(8)
2
6 mount.gfs2 - GFS2 mount options
7
10 mount -a [-fnrsvw] -t gfs2 [-O options]
11 mount [-fnrsvw] -t gfs2 [-o options ] device dir
12
15 For details on the common mount options, please see the mount(8) com‐
16 mand man page. The device may be any block device on which you have
17 created a GFS2 filesystem. Examples include a single disk partition
18 (e.g. /dev/sdb3), a loopback device, a device exported from another
19 node (e.g. an iSCSI device), or a logical volume (typically comprised
20 of a number of individual disks).
21 device does not necessarily need to match the device name as seen on
23 another node in the cluster, nor does it need to be a logical volume.
24 However, the use of a cluster-aware volume manager such as CLVM2 (see
25 lvm(8)) will guarantee that the managed devices are named identically
26 on each node in a cluster (for much easier management), and will allow
27 you to configure a very large volume from multiple storage units (e.g.
28 disk drives).
29 device must make the entire filesystem storage area visible to the com‐
31 puter. That is, you cannot mount different parts of a single filesys‐
32 tem on different computers. Each computer must see an entire filesys‐
33 tem. You may, however, mount several GFS2 filesystems if you want to
34 distribute your data storage in a controllable way.
35 This man page describes GFS2-specific options that can be passed to the
37 GFS2 file system at mount time, using the -o flag. There are many
38 other -o options handled by the generic mount command mount(8). How‐
39 ever, the options described below are specifically for GFS2, and are
40 not interpreted by the mount command nor by the kernel's Virtual File
41 System. GFS2 and non-GFS2 options may be intermingled after the -o,
42 separated by commas (but no spaces).
43 The options commit, discard, errors, quota_quantum, statfs_quantum,
45 statfs_percent, barrier, acl, quota, suiddir, and data can be changed
46 after mount using the "mount -o remount,option /mountpoint" command.
47 The options quota, discard, barrier, acl, and suiddir support the "no"
48 prefix. For example, "noacl" turns off what "acl" turns on.
49 If you have trouble mounting GFS2, check the syslog (e.g. /var/log/mes‐
51 sages) for specific error messages.
52
55 lockproto=LockProtoName
56 This specifies which inter-node lock protocol is used by the
57 GFS2 filesystem for this mount, overriding the default lock pro‐
58 tocol name stored in the filesystem's on-disk superblock.
59 The LockProtoName must be one of the supported locking proto‐
61 cols, currently these are lock_nolock and lock_dlm.
62 The default lock protocol name is written to disk initially when
64 creating the filesystem with mkfs.gfs2(8), -p option. It can be
65 changed on-disk by using the gfs2_tool(8) utility's sb proto
66 command.
67 The lockproto mount option should be used only under special
69 circumstances in which you want to temporarily use a different
70 lock protocol without changing the on-disk default. Using the
71 incorrect lock protocol on a cluster filesystem mounted from
72 more than one node will almost certainly result in filesystem
73 corruption.
74 locktable=LockTableName
76 This specifies the identity of the cluster and of the filesystem
77 for this mount, overriding the default cluster/filesystem iden‐
78 tify stored in the filesystem's on-disk superblock. The clus‐
79 ter/filesystem name is recognized globally throughout the clus‐
80 ter, and establishes a unique namespace for the inter-node lock‐
81 ing system, enabling the mounting of multiple GFS2 filesystems.
82 The format of LockTableName is lock-module-specific. For
84 lock_dlm, the format is clustername:fsname. For lock_nolock,
85 the field is ignored.
86 The default cluster/filesystem name is written to disk initially
88 when creating the filesystem with mkfs.gfs2(8), -t option. It
89 can be changed on-disk by using the gfs2_tool(8) utility's sb
90 table command.
91 The locktable mount option should be used only under special
93 circumstances in which you want to mount the filesystem in a
94 different cluster, or mount it as a different filesystem name,
95 without changing the on-disk default.
96 localcaching
98 This flag tells GFS2 that it is running as a local (not clus‐
99 tered) filesystem, so it can turn on some block caching opti‐
100 mizations that can't be used when running in cluster mode.
101 This is turned on automatically by the lock_nolock module, but
103 can be overridden by using the ignore_local_fs option.
104 localflocks
106 This flag tells GFS2 that it is running as a local (not clus‐
107 tered) filesystem, so it can allow the kernel VFS layer to do
108 all flock and fcntl file locking. When running in cluster mode,
109 these file locks require inter-node locks, and require the sup‐
110 port of GFS2. When running locally, better performance is
111 achieved by letting VFS handle the whole job.
112 This is turned on automatically by the lock_nolock module, but
114 can be overridden by using the ignore_local_fs option.
115 errors=[panic|withdraw]
117 Setting errors=panic causes GFS2 to oops when encountering an
118 error that would otherwise cause the mount to withdraw or print
119 an assertion warning. The default setting is errors=withdraw.
120 This option should not be used in a production system. It
121 replaces the earlier debug option on kernel versions 2.6.31 and
122 above.
123 ignore_local_fs
125 By default, using the nolock lock module automatically turns on
126 the localcaching and localflocks optimizations. ignore_local_fs
127 forces GFS2 to treat the filesystem as if it were a multihost
128 (clustered) filesystem, with localcaching and localflocks opti‐
129 mizations turned off.
130 upgrade
132 This flag tells GFS2 to upgrade the filesystem's on-disk format
133 to the version supported by the current GFS2 software installa‐
134 tion on this computer. If you try to mount an old-version disk
135 image, GFS2 will notify you via a syslog message that you need
136 to upgrade. Try mounting again, using the -o upgrade option.
137 When upgrading, only one node may mount the GFS2 filesystem.
138 acl Enables POSIX Access Control List acl(5) support within GFS2.
140 spectator
142 Mount this filesystem using a special form of read-only mount.
143 The mount does not use one of the filesystem's journals. The
144 node is unable to recover journals for other nodes.
145 suiddir
147 Sets owner of any newly created file or directory to be that of
148 parent directory, if parent directory has S_ISUID permission
149 attribute bit set. Sets S_ISUID in any new directory, if its
150 parent directory's S_ISUID is set. Strips all execution bits on
151 a new file, if parent directory owner is different from owner of
152 process creating the file. Set this option only if you know why
153 you are setting it.
154 quota=[off/account/on]
156 Turns quotas on or off for a filesystem. Setting the quotas to
157 be in the "account" state causes the per UID/GID usage statis‐
158 tics to be correctly maintained by the filesystem, limit and
159 warn values are ignored. The default value is "off".
160 discard
162 Causes GFS2 to generate "discard" I/O requests for blocks which
163 have been freed. These can be used by suitable hardware to
164 implement thin-provisioning and similar schemes. This feature is
165 supported in kernel version 2.6.30 and above.
166 barrier
168 This option, which defaults to on, causes GFS2 to send I/O bar‐
169 riers when flushing the journal. The option is automatically
170 turned off if the underlying device does not support I/O barri‐
171 ers. We highly recommend the use of I/O barriers with GFS2 at
172 all times unless the block device is designed so that it cannot
173 lose its write cache content (e.g. its on a UPS, or it doesn't
174 have a write cache)
175 commit=secs
177 This is similar to the ext3 commit= option in that it sets the
178 maximum number of seconds between journal commits if there is
179 dirty data in the journal. The default is 60 seconds. This
180 option is only provided in kernel versions 2.6.31 and above.
181 data=[ordered|writeback]
183 When data=ordered is set, the user data modified by a transac‐
184 tion is flushed to the disk before the transaction is committed
185 to disk. This should prevent the user from seeing uninitialized
186 blocks in a file after a crash. Data=writeback mode writes the
187 user data to the disk at any time after it's dirtied. This
188 doesn't provide the same consistency guarantee as ordered mode,
189 but it should be slightly faster for some workloads. The
190 default is ordered mode.
191 meta This option results in selecting the meta filesystem root rather
193 than the normal filesystem root. This option is normally only
194 used by the GFS2 utility functions. Altering any file on the
195 GFS2 meta filesystem may render the filesystem unusable, so only
196 experts in the GFS2 on-disk layout should use this option.
197 quota_quantum=secs
199 This sets the number of seconds for which a change in the quota
200 information may sit on one node before being written to the
201 quota file. This is the preferred way to set this parameter. The
202 value is an integer number of seconds greater than zero. The
203 default is 60 seconds. Shorter settings result in faster updates
204 of the lazy quota information and less likelihood of someone
205 exceeding their quota. Longer settings make filesystem opera‐
206 tions involving quotas faster and more efficient.
207 statfs_quantum=secs
209 Setting statfs_quantum to 0 is the preferred way to set the slow
210 version of statfs. The default value is 30 secs which sets the
211 maximum time period before statfs changes will be syned to the
212 master statfs file. This can be adjusted to allow for faster,
213 less accurate statfs values or slower more accurate values. When
214 set to 0, statfs will always report the true values.
215 statfs_percent=value
217 This setting provides a bound on the maximum percentage change
218 in the statfs information on a local basis before it is synced
219 back to the master statfs file, even if the time period has not
220 expired. If the setting of statfs_quantum is 0, then this set‐
221 ting is ignored.
222
225 GFS2 doesn't support errors=remount-ro or data=journal. It is not pos‐
226 sible to switch support for user and group quotas on and off indepen‐
227 dently of each other. Some of the error messages are rather cryptic, if
228 you encounter one of these messages check firstly that gfs_controld is
229 running and secondly that you have enough journals on the filesystem
230 for the number of nodes in use.
231
234 gfs2(8), mount(8) for general mount options, chmod(1) and chmod(2) for
235 access permission flags, acl(5) for access control lists, lvm(8) for
236 volume management, ccs(7) for cluster management, umount(8), initrd(4).
237 mount.gfs2(8)