1VFS_GPFS(8) System Administration tools VFS_GPFS(8)
2
6 vfs_gpfs - gpfs specific samba extensions like acls and prealloc
7
9 vfs objects = gpfs
10
12 This VFS module is part of the samba(7) suite.
13 The gpfs VFS module is the home for all gpfs extensions that Samba
15 requires for proper integration with GPFS. It uses the GPL library
16 interfaces provided by GPFS.
17 Currently the gpfs vfs module provides extensions in following areas :
19 · NFSv4 ACL Interfaces with configurable options for GPFS
21 · Kernel oplock support on GPFS
23 · Lease support on GPFS
25 NOTE: This module follows the posix-acl behaviour and hence allows
28 permission stealing via chown. Samba might allow at a later point in
29 time, to restrict the chown via this module as such restrictions are
30 the responsibility of the underlying filesystem than of Samba.
31 This module makes use of the smb.conf parameter acl map full control.
33 When set to yes (the default), this parameter will add in the
34 FILE_DELETE_CHILD bit on a returned ACE entry for a file (not a
35 directory) that already contains all file permissions except for
36 FILE_DELETE and FILE_DELETE_CHILD. This can prevent Windows
37 applications that request GENERIC_ALL access from getting ACCESS_DENIED
38 errors when running against a filesystem with NFSv4 compatible ACLs.
39 This module is stackable.
41 Since Samba 4.0 all options are per share options.
43
45 gpfs:sharemodes = [ yes | no ]
46 Enable/Disable cross node sharemode handling for GPFS.
47 · yes(default) - propagate sharemodes across all GPFS
49 nodes.
50 · no - do not propagate sharemodes across all GPFS nodes.
52 This should only be used if the GPFS file system is
53 exclusively exported by Samba. Access by local unix
54 application or NFS exports could lead to corrupted
55 files.
56 gpfs:leases = [ yes | no ]
59 Enable/Disable cross node leases (oplocks) for GPFS. You should
60 also set the oplocks and kernel oplocks options to the same value.
61 · yes(default) - propagate leases across all GPFS nodes.
63 · no - do not propagate leases across all GPFS nodes. This
65 should only be used if the GPFS file system is
66 exclusively exported by Samba. Access by local unix
67 application or NFS exports could lead to corrupted
68 files.
69 gpfs:hsm = [ yes | no ]
72 Enable/Disable announcing if this FS has HSM enabled.
73 · no(default) - Do not announce HSM.
75 · yes - Announce HSM.
77 gpfs:recalls = [ yes | no ]
80 When this option is set to no, an attempt to open an offline file
81 will be rejected with access denied. This helps preventing recall
82 storms triggered by careless applications like Finder and Explorer.
83 · yes(default) - Open files that are offline. This will
85 recall the files from HSM.
86 · no - Reject access to offline files with access denied.
88 This will prevent recalls of files from HSM. Using this
89 setting also requires gpfs:hsm to be set to yes.
90 gpfs:getrealfilename = [ yes | no ]
93 Enable/Disable usage of the gpfs_get_realfilename_path() function.
94 This improves the casesensitive wildcard file name access.
95 · yes(default) - use gpfs_get_realfilename_path().
97 · no - do not use gpfs_get_realfilename_path(). It seems
99 that gpfs_get_realfilename_path() doesn't work on AIX.
100 gpfs:winattr = [ yes | no ]
103 Enable/Disable usage of the windows attributes in GPFS. GPFS is
104 able to store windows file attributes e.g. HIDDEN, READONLY, SYSTEM
105 and others natively. That means Samba doesn't need to map them to
106 permission bits or extended attributes.
107 · no(default) - do not use GPFS windows attributes.
109 · yes - use GPFS windows attributes.
111 gpfs:merge_writeappend = [ yes | no ]
114 GPFS ACLs doesn't know about the 'APPEND' right. This option lets
115 Samba map the 'APPEND' right to 'WRITE'.
116 · yes(default) - map 'APPEND' to 'WRITE'.
118 · no - do not map 'APPEND' to 'WRITE'.
120 gpfs:acl = [ yes | no ]
123 This option lets Samba use or ignore GPFS ACLs.
124 · yes(default) - use GPFS ACLs.
126 · no - do not use GPFS ACLs and pass everything to the
128 next SMB_VFS module.
129 gpfs:refuse_dacl_protected = [ yes | no ]
132 As GPFS does not support the ACE4_FLAG_NO_PROPAGATE NFSv4 flag
133 (which would be the mapping for the DESC_DACL_PROTECTED flag), the
134 status of this flag is currently silently ignored by Samba. That
135 means that if you deselect the "Allow inheritable permissions..."
136 checkbox in Windows' ACL dialog and then apply the ACL, the flag
137 will be back immediately.
138 To make sure that automatic migration with e.g. robocopy does not
140 lead to ACLs silently (and unintentionally) changed, you can set
141 gpfs:refuse_dacl_protected = yes to enable an explicit check for
142 this flag and if set, it will return NT_STATUS_NOT_SUPPORTED so
143 errors are shown up on the Windows side and the Administrator is
144 aware of the ACLs not being settable like intended
145 · no(default) - ignore the DESC_DACL_PROTECTED flags.
147 · yes - reject ACLs with DESC_DACL_PROTECTED.
149 gpfs:dfreequota = [ yes | no ]
152 Adjust reporting of the size and free space of a share according to
153 quotas. If this setting is "yes", a request for size and free space
154 will also evaluate the user quota of the user requesting the data
155 and the group quota of the primary group of the user. Fileset
156 quotas are not queried, since GPFS already provides the option
157 --dfreequota to reflect the fileset quota in the free space query.
158 Please use that option to include fileset quotas in the reported
159 disk space.
160 If any of the soft or hard quota limits has been reached, the free
162 space will be reported as 0. If a quota is in place, but the limits
163 have not been reached, the free space will be reported according to
164 the space left in the quota. If more than one quota applies the
165 free space will be reported as the smallest space left in those
166 quotas. The size of the share will be reported according to the
167 quota usage. If more than one quota applies, the smallest size will
168 be reported for the share size according to these quotas.
169 · yes - include the quotas when reporting the share size
171 and free space
172 · no(default) - do not include quotas, simply report the
174 size and free space of the file system
175 gpfs:prealloc = [ yes | no ]
178 If set to yes the gpfs_prealloc function will be used in the
179 fallocate callback when appropriate. If set to no gpfs_prealloc
180 will not be used. In both cases the system and libc calls are
181 avoided.
182 · yes (default) - Use gpfs_prealloc for the fallocate
184 callback.
185 · no - Do not use gpfs_prealloc for the fallocate
187 callback.
188 gpfs:settimes = [ yes | no ]
191 Use the gpfs_set_times API when changing the timestamps of a file
192 or directory. If the GPFS API is not available the old method of
193 using utime and the GPFS winattr call will be used instead.
194 · yes(default) - Use gpfs_set_times. Fall back to utime
196 and winattr when it is not available.
197 · no - Do not use gpfs_set_times.
199 nfs4:mode = [ simple | special ]
202 Controls substitution of special IDs (OWNER@ and GROUP@) on GPFS.
203 The use of mode simple is recommended. In this mode only non
204 inheriting ACL entries for the file owner and group are mapped to
205 special IDs.
206 The following MODEs are understood by the module:
208 · simple(default) - use OWNER@ and GROUP@ special IDs for
210 non inheriting ACEs only.
211 · special(deprecated) - use OWNER@ and GROUP@ special IDs
213 in ACEs for all file owner and group ACEs.
214 nfs4:acedup = [dontcare|reject|ignore|merge]
217 This parameter configures how Samba handles duplicate ACEs
218 encountered in GPFS ACLs. GPFS allows/creates duplicate ACE for
219 different bits for same ID.
220 Following is the behaviour of Samba for different values :
222 · dontcare (default) - copy the ACEs as they come
224 · reject - stop operation and exit with error on ACL set
226 op
227 · ignore - don't include the second matching ACE
229 · merge - bitwise OR the 2 ace.flag fields and 2 ace.mask
231 fields of the 2 duplicate ACEs into 1 ACE
232 nfs4:chown = [yes|no]
235 This parameter allows enabling or disabling the chown supported by
236 the underlying filesystem. This parameter should be enabled with
237 care as it might leave your system insecure.
238 Some filesystems allow chown as a) giving b) stealing. It is the
240 latter that is considered a risk.
241 Following is the behaviour of Samba for different values :
243 · yes - Enable chown if as supported by the under
245 filesystem
246 · no (default) - Disable chown
248 gpfs:syncio = [yes|no]
251 This parameter makes Samba open all files with O_SYNC. This
252 triggers optimizations in GPFS for workloads that heavily share
253 files.
254 Following is the behaviour of Samba for different values:
256 · yes - Open files with O_SYNC
258 · no (default) - Open files as normal Samba would do
260
263 A GPFS mount can be exported via Samba as follows :
264 [samba_gpfs_share]
266 vfs objects = gpfs
267 path = /test/gpfs_mount
268 nfs4: mode = special
269 nfs4: acedup = merge
270
272 Depending on the version of gpfs, the libgpfs_gpl library or the
273 libgpfs library is needed at runtime by the gpfs VFS module: Starting
274 with gpfs 3.2.1 PTF8, the complete libgpfs is available as open source
275 and libgpfs_gpl does no longer exist. With earlier versions of gpfs,
276 only the libgpfs_gpl library was open source and could be used at run
277 time.
278 At build time, only the header file gpfs_gpl.h is required, which is a
280 symlink to gpfs.h in gpfs versions newer than 3.2.1 PTF8.
281
283 This man page is part of version 4.9.1 of the Samba suite.
284
286 The original Samba software and related utilities were created by
287 Andrew Tridgell. Samba is now developed by the Samba Team as an Open
288 Source project similar to the way the Linux kernel is developed.
289 The GPFS VFS module was created with contributions from Volker Lendecke
291 and the developers at IBM.
292 This manpage was created by the IBM FSCC team
294Samba 4.9.1 05/11/2019 VFS_GPFS(8)