1
2FEDFS-MAP-NFS4(8)           System Manager's Manual          FEDFS-MAP-NFS4(8)
3
4
5

NAME

7       fedfs-map-nfs4 - generate automounter program map entries for FedFS
8

SYNOPSIS

10       fedfs-map-nfs4 domainname
11

INTRODUCTION

13       RFC  5716  introduces  the  Federated  File  System (FedFS, for short).
14       FedFS is an extensible standardized mechanism by which system  adminis‐
15       trators  construct  a  coherent  namespace across multiple file servers
16       using file system referrals.  For further details, see fedfs(7).
17

DESCRIPTION

19       The fedfs-map-nfs4(8) command provides a  FedFS  program  map  for  the
20       local  system's  automounter.   Although it is typically intended to be
21       invoked by the automounter, it is also  safe  to  invoke  directly  for
22       scripting  or  debugging purposes.  See autofs(5) for information about
23       how program maps work.
24
25   Operation
26       The fedfs-map-nfs4(8) command locates FedFS domains by looking for  DNS
27       SRV  records  that  advertise  file servers exporting FedFS domain root
28       replicas.  The domainname argument determines what FedFS domain  is  to
29       be mounted.
30
31       It retrieves and sorts the domain root replica records according to SRV
32       record sorting rules outlined in RFC 2782.  It  then  generates  a  sun
33       format map entry on stdout representing the set of servers contained in
34       the SRV record, a standard export path to the domain root,  and  appro‐
35       priate NFS mount options.  Error messages are output on stderr.
36
37   Globally useful names
38       Across all FedFS-enabled file system clients, a unique file object in a
39       FedFS domain is always accessed via the same pathname.  Such  pathnames
40       are referred to as globally useful names.  See fedfs(7) for a full dis‐
41       cussion.
42
43       The top-level directory of a globally useful name is  always  the  net‐
44       worked file system type (NFS version 4, CIFS, and so on).  A fedfs-map-
45       nfs4(8) program map entry is used with  the  NFS  version  4  top-level
46       directory to provide globally useful names via the NFS version 4 proto‐
47       col.
48

EXAMPLES

50       Typically, a fedfs-map-nfs4(8) entry  in  /etc/auto.master  looks  like
51       this:
52
53              /nfs4  /usr/sbin/fedfs-map-nfs4
54
55       Under  the  /nfs4  directory  on the local system, the automounter uses
56       fedfs-map-nfs4(8) to convert a FedFS domain name to a  set  of  servers
57       and  an  export path, which are then passed to mount.nfs(8).  The auto‐
58       mounter mounts this FedFS domain on the directory /nfs4/domainname.
59
60       After configuring and restarting autofs, to access files in  the  exam‐
61       ple.net FedFS domain, for instance, you can start with:
62
63              $ cd /nfs4/example.net
64
65       The  automounter uses the fedfs-map-nfs4(8) command to look up the file
66       servers that provide the domain root for the  example.net  domain.   It
67       then mounts one of these servers on /nfs4/example.net.
68
69       If  the  fedfs-map-nfs4(8) command cannot find the requested domain, no
70       local directory is created and no mount operation is performed.  Appli‐
71       cations receive an ENOENT error in this case.
72
73       While  these  mounted  domains  remain  active on the local system, the
74       mounted-on directories remain visible.  After a period  of  inactivity,
75       the automounter automatically unmounts a FedFS domain.
76
77       Local  applications  browsing  the  top-level  directory do not see all
78       available FedFS domains.  They see only the ones that are  mounted  and
79       active.
80
81   Mount option inheritance
82       The Linux NFS client treats an NFS referral as a server-initiated mount
83       request.  The referring fileserver provides only a list of server names
84       and  export  paths.  The mount options for this new mount are inherited
85       from the new mount point's parent directory on the client.
86
87       As applications proceed deeper into  a  domain's  namespace,  they  can
88       encounter  both file sets to which they have read-only access, and file
89       sets to which they  have  read-write  access.   To  allow  applications
90       proper access to both types of file sets, typically file-access clients
91       mount domain root directories in read-write mode.  All submounts of the
92       domain  root are then mounted read-write as well.  Write access is con‐
93       trolled by fileservers.
94
95       For example, a domain root may contain an NFS version 4 referral to  an
96       export  containing  user  home  directories.   The  domain  root may be
97       exported read-only so file-access clients cannot update  it,  but  user
98       home  directories would not be very useful if they could not be written
99       to by their owners.  The fileserver continues to  employ  user  creden‐
100       tials to limit access as appropriate.
101
102       Network  file  system  clients follow file system referrals as applica‐
103       tions encounter them, which is similar to  how  an  automounter  works.
104       Consider  the initial mount of the domain root as if you are mounting a
105       single whole file system, even though underneath, additional NFS mounts
106       come and go as needed.
107

FILES

109       /etc/auto.master  master automounter map
110

SEE ALSO

112       fedfs(7), nfs(5), autofs(5),
113
114       RFC 2782 for a discussion of DNS SRV records
115
116       RFC 5716 for FedFS requirements and overview
117

COLOPHON

119       This  page  is  part  of the fedfs-utils package.  A description of the
120       project  and  information  about  reporting  bugs  can  be   found   at
121       http://wiki.linux-nfs.org/wiki/index.php/FedFsUtilsProject.
122

AUTHOR

124       Chuck Lever <chuck.lever@oracle.com>
125
126
127
128                                3 February 2014              FEDFS-MAP-NFS4(8)
Impressum