vfs_fileid — Generates file_id structs with unique device id values for cluster setups. It also adds ways to deliberately break lock coherency for specific inodes
vfs objects = fileid
This VFS module is part of the samba(7) suite.
Samba uses file_id structs to uniquely identify files
for locking purpose. By default the file_id contains the device
and inode number returned by the stat()
system call.
As the file_id is a unique identifier of a file, it must be the same
on all nodes in a cluster setup. This module overloads the
SMB_VFS_FILE_ID_CREATE()
operation and
generates the device number based on the configured algorithm
(see the "fileid:algorithm" option).
When using the fsname or fsid algorithm a
stat()
and statfs()
call is
required for all mounted file systems to generate the file_id. If e.g.
an NFS file system is unresponsive such a call might block and the smbd
process will become unresponsive. Use the "fileid:fstype deny",
"fileid:fstype allow", "fileid:mntdir deny", or "fileid:mntdir allow"
options to ignore potentially unresponsive file systems.
Available algorithms are fsname
,
fsid
, next_module
. The default value is
fsname
. As well as the following legacy
algorithms: fsname_nodirs
, fsname_norootdir
,
fsname_norootdir_ext
and hostname
.
The fsname
algorithm generates
device id by hashing the kernel device name.
The fsid
algorithm generates
the device id from the f_fsid
returned
from the statfs()
syscall.
The next_module
algorithm lets the next vfs module
in the module chain generate the id. This is mainly used in combination
with the various 'nolock' features the fileid module provides.
The legacy hostname
algorithm generates unique
devid by hashing the hostname and low level device id.
It also implies fileid:nolock_all_inodes=yes
.
This can be used to deliberately break lock coherency in a cluster
and with fileid:nolock_max_slots
also between local processes
within a node. NOTE: Do not use this without knowing what you are doing!
It breaks SMB semantics and it can lead to data corruption!
This implies fileid:nolock_all_inodes=yes
.
The legacy fsname_nodirs
algorithm is an alias
for using the fsname
algorithm together with
fileid:nolock_all_dirs=yes
.
NOTE: Do not use this without knowing what you are doing!
It breaks SMB semantics!
See fileid:nolock_paths
for a more fine grained
approach.
The legacy fsname_norootdir
algorithm is an alias
for using the fsname
algorithm together with
fileid:nolock_paths= “.”
. It means
this can be used to deliberately break lock coherency
in a cluster for the root directory of a share.
The legacy fsname_norootdir_ext
algorithm is an alias
for using the fsname
algorithm together with
fileid:nolock_paths= “.”
and
fileid:nolock_max_slots = 18446744073709551615
.
It means this can be used to deliberately break lock coherency
completely for the root directory of a share. Even local processes
are no longer lock coherent.
This option is the legacy version of the
fileid:algorithm
option, which was used in earlier
versions of fileid mapping feature in custom Samba 3.0 versions.
List of file system types to be ignored for file_id generation.
List of file system types to be allowed for file_id generation. If this option is set, file system types not listed here are ignored.
List of file system mount points to be ignored for file_id generation.
List of file system mount points to be allowed for file_id generation. If this option is set, file system mount points not listed here are ignored.
This option alters the behavior of the nolock
algorithm
in a ways that it also breaks the lock coherency between individual processes
on the same host. The default is to have just 1 concurrent slot available per host.
By incressing the number of slots you can specify how many concurrent processes
can work on a given inode without contention, the number should typically be larger
than the a number of logical cpus, maybe 2 times of num_cpus.
This option triggers the use of the fileid nolock behavior
for all directory inodes, which can be used to deliberately break
the lock coherency for all directories.
NOTE: Do not use this without knowing what you are doing!
It breaks SMB semantics!
See fileid:nolock_paths
for a more fine grained
approach.
This option triggers the use of the fileid nolock algorithm
for all directoriy inode, which can be used to deliberately break
the lock coherency for all directories.
NOTE: Do not use this without knowing what you are doing!
It breaks SMB semantics and it can lead to data corruption!
See fileid:nolock_paths
for a more fine grained
approach.
This option specifies a path list referring to files and/or directories, which should use fileid nolock algorithm in order to deliberately break the lock coherency for them. The specified paths can be relative to the share root directory or absolute. The names are case sensitive unix pathnames! Note all paths are only evaluated at tree connect time, when the share is being connected, from there on only the related device and inode numbers from the stat() syscall are compared. Non existing paths will generate a log level 0 message. NOTE: This option should be used with care as it breaks SMB semantics! But it may help in situation where a specific (commonly read-only) inode is highly contended.
This legacy option triggers use of the fileid nolock behavior
for the configured inode, while ignoring and device id. This can be used to deliberately break
lock coherency for the corresponding file or directory in a
cluster. Using the fileid:nolock_paths
option is much more flexible and simpler to use.
Usage of the fileid
module with the
fsid
algorithm:
[global]
vfs objects = fileid
fileid:algorithm = fsid
Usage of the fileid
module in order
avoid load on heavily contended (most likely read-only) inodes.
[global]
vfs objects = fileid
fileid:algorithm = next_module
fileid:nolock_paths = . ContendedFolder1 /path/to/contended.exe
fileid:nolock_max_slots = 256