\input texinfo @c -*-texinfo-*- @c @c Copyright (c) 1989 Jan-Simon Pendry @c Copyright (c) 1989 Imperial College of Science, Technology & Medicine @c Copyright (c) 1989 The Regents of the University of California. @c All rights reserved. @c @c This code is derived from software contributed to Berkeley by @c Jan-Simon Pendry at Imperial College, London. @c @c Redistribution and use in source and binary forms, with or without @c modification, are permitted provided that the following conditions @c are met: @c 1. Redistributions of source code must retain the above copyright @c notice, this list of conditions and the following disclaimer. @c 2. Redistributions in binary form must reproduce the above copyright @c notice, this list of conditions and the following disclaimer in the @c documentation and/or other materials provided with the distribution. @c 3. All advertising materials mentioning features or use of this software @c must display the following acknowledgement: @c This product includes software developed by the University of @c California, Berkeley and its contributors. @c 4. Neither the name of the University nor the names of its contributors @c may be used to endorse or promote products derived from this software @c without specific prior written permission. @c @c THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND @c ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE @c IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE @c ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE @c FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL @c DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS @c OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) @c HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT @c LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY @c OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF @c @c from: @(#)amdref.texinfo 8.1 (Berkeley) 6/6/93 @c $Id: amdref.texinfo,v 1.11 2004/02/19 23:40:38 deraadt Exp $ @c @setfilename amdref.info @c @setfilename /usr/local/emacs/info/amd @tex \overfullrule=0pt @end tex @settitle 4.4 BSD Automounter Reference Manual @dircategory Miscellaneous @direntry * Amdref: (amdref). 4.4 BSD Automounter Reference Manual @end direntry @titlepage @sp 6 @center @titlefont{Amd} @sp 2 @center @titlefont{The 4.4 BSD Automounter} @sp 2 @center @titlefont{Reference Manual} @sp 2 @center @authorfont{Jan-Simon Pendry} @sp @center @i{and} @sp @center @authorfont{Nick Williams} @sp 4 @center Last updated March 1991 @center Documentation for software revision 5.3 Alpha @page Copyright @copyright{} 1989 Jan-Simon Pendry @sp -1 Copyright @copyright{} 1989 Imperial College of Science, Technology & Medicine @sp -1 Copyright @copyright{} 1989 The Regents of the University of California. @sp 0 All Rights Reserved. @vskip 1ex Permission to copy this document, or any portion of it, as necessary for use of this software is granted provided this copyright notice and statement of permission are included. @end titlepage @page @ifinfo @node Top, License, , (DIR) Amd - The 4.4 BSD Automounter ***************************** Amd is the 4.4 BSD Automounter. This Info file describes how to use and understand Amd. @end ifinfo @menu * License:: Explains the terms and conditions for using and distributing Amd. * Distrib:: How to get the latest Amd distribution. * Intro:: An introduction to Automounting concepts. * Overview:: An overview of Amd. * Supported Platforms:: Machines and Systems supported by Amd. * Mount Maps:: Details of mount maps * Amd Command Line Options:: All the Amd command line options explained. * Filesystem Types:: The different mount types supported by Amd. * Run-time Administration:: How to start, stop and control Amd. * Examples:: Some examples showing how Amd might be used. * Internals:: Implementation details. * Acknowledgements & Trademarks:: Legal notes. Indexes * Index:: An item for each concept. @end menu @iftex @unnumbered Preface This manual documents the use of the 4.4 BSD automounter---@i{Amd}. This is primarily a reference manual. Unfortunately, no tutorial exists. This manual comes in two forms: the published form and the Info form. The Info form is for on-line perusal with the INFO program which is distributed along with GNU Emacs. Both forms contain substantially the same text and are generated from a common source file, which is distributed with the @i{Amd} source. @end iftex @node License, Distrib, Top, Top @unnumbered License @cindex License Information @i{Amd} is not in the public domain; it is copyrighted and there are restrictions on its distribution. Redistribution and use in source and binary forms are permitted provided that: (1) source distributions retain this entire copyright notice and comment, and (2) distributions including binaries display the following acknowledgement: ``This product includes software developed by The University of California, Berkeley and its Contributors'' in the documentation or other materials provided with the distribution and in all advertising materials mentioning features or use of this software. neither the name of the University nor the names of its Contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. @node Distrib, Intro, License, Top @unnumbered Source Distribution @cindex Source code distribution @cindex Obtaining the source code If you have access to the Internet, you can get the latest distribution version of @i{Amd} from host @file{usc.edu} using anonymous FTP. Move to the directory @file{/pub/amd} on that host and fetch the file @file{amd.tar.Z}. If you are in the UK, you can get the latest distribution version of @i{Amd} from the UKnet info-server. Start by sending email to @file{info-server@@doc.ic.ac.uk}. Sites on the UK JANET network can get the latest distribution by using anonymous NIFTP to fetch the file @samp{amd.tar.Z} from host @samp{uk.ac.imperial.doc.src}. Revision 5.2 was part of the 4.3 BSD Reno distribution. Revision 5.3bsdnet, a late alpha version of 5.3, was part of the BSD network version 2 distribution @unnumberedsec Bug Reports @cindex Bug reports Send all bug reports to @file{jsp@@doc.ic.ac.uk} quoting the details of the release and your configuration. These can be obtained by running the command @samp{amd -v}. @unnumberedsec Mailing List @cindex Mailing list There is a mailing list for people interested in keeping uptodate with developments. To subscribe, send a note to @file{amd-workers-request@@acl.lanl.gov}. @node Intro, Overview, Distrib, Top @unnumbered Introduction @cindex Introduction An @dfn{automounter} maintains a cache of mounted filesystems. Filesystems are mounted on demand when they are first referenced, and unmounted after a period of inactivity. @i{Amd} may be used as a replacement for Sun's automounter. The choice of which filesystem to mount can be controlled dynamically with @dfn{selectors}. Selectors allow decisions of the form ``hostname is @var{this},'' or ``architecture is not @var{that}.'' Selectors may be combined arbitrarily. @i{Amd} also supports a variety of filesystem types, including NFS, UFS and the novel @dfn{program} filesystem. The combination of selectors and multiple filesystem types allows identical configuration files to be used on all machines so reducing the administrative overhead. @i{Amd} ensures that it will not hang if a remote server goes down. Moreover, @i{Amd} can determine when a remote server has become inaccessible and then mount replacement filesystems as and when they become available. @i{Amd} contains no proprietary source code and has been ported to numerous flavours of Unix. @node Overview, Supported Platforms, Intro, Top @chapter Overview @i{Amd} maintains a cache of mounted filesystems. Filesystems are @dfn{demand-mounted} when they are first referenced, and unmounted after a period of inactivity. @i{Amd} may be used as a replacement for Sun's @b{automount}(8) program. It contains no proprietary source code and has been ported to numerous flavours of Unix. @xref{Supported Operating Systems}.@refill @i{Amd} was designed as the basis for experimenting with filesystem layout and management. Although @i{Amd} has many direct applications it is loaded with additional features which have little practical use. At some point the infrequently used components may be removed to streamline the production system. @c @i{Amd} supports the notion of @dfn{replicated} filesystems by evaluating @c each member of a list of possible filesystem locations in parallel. @c @i{Amd} checks that each cached mapping remains valid. Should a mapping be @c lost -- such as happens when a fileserver goes down -- @i{Amd} automatically @c selects a replacement should one be available. @c @menu * Fundamentals:: * Filesystems and Volumes:: * Volume Naming:: * Volume Binding:: * Operational Principles:: * Mounting a Volume:: * Automatic Unmounting:: * Keep-alives:: * Non-blocking Operation:: @end menu @node Fundamentals, Filesystems and Volumes, Overview, Overview @comment node-name, next, previous, up @section Fundamentals @cindex Automounter fundamentals The fundamental concept behind @i{Amd} is the ability to separate the name used to refer to a file from the name used to refer to its physical storage location. This allows the same files to be accessed with the same name regardless of where in the network the name is used. This is very different from placing @file{/n/hostname} in front of the pathname since that includes location dependent information which may change if files are moved to another machine. By placing the required mappings in a centrally administered database, filesystems can be re-organised without requiring changes to configuration files, shell scripts and so on. @node Filesystems and Volumes, Volume Naming, Fundamentals, Overview @comment node-name, next, previous, up @section Filesystems and Volumes @cindex Filesystem @cindex Volume @cindex Fileserver @cindex sublink @i{Amd} views the world as a set of fileservers, each containg one or more filesystems where each filesystem contains one or more @dfn{volumes}. Here the term @dfn{volume} is used to refer to a coherent set of files such as a user's home directory or a @TeX{} distribution.@refill In order to access the contents of a volume, @i{Amd} must be told in which filesystem the volume resides and which host owns the filesystem. By default the host is assumed to be local and the volume is assumed to be the entire filesystem. If a filesystem contains more than one volume, then a @dfn{sublink} is used to refer to the sub-directory within the filesystem where the volume can be found. @node Volume Naming, Volume Binding, Filesystems and Volumes, Overview @comment node-name, next, previous, up @section Volume Naming @cindex Volume names @cindex Network-wide naming @cindex Replicated volumes @cindex Duplicated volumes @cindex Replacement volumes Volume names are defined to be unique across the entire network. A volume name is the pathname to the volume's root as known by the users of that volume. Since this name uniquely identifies the volume contents, all volumes can be named and accessed from each host, subject to administrative controls. Volumes may be replicated or duplicated. Replicated volumes contain identical copies of the same data and reside at two or more locations in the network. Each of the replicated volumes can be used interchangeably. Duplicated volumes each have the same name but contain different, though functionally identical, data. For example, @samp{/vol/tex} might be the name of a @TeX{} distribution which varied for each machine architecture.@refill @i{Amd} provides facilities to take advantage of both replicated and duplicated volumes. Configuration options allow a single set of configuration data to be shared across an entire network by taking advantage of replicated and duplicated volumes. @i{Amd} can take advantage of replacement volumes by mounting them as required should an active fileserver become unavailable. @node Volume Binding, Operational Principles, Volume Naming, Overview @comment node-name, next, previous, up @section Volume Binding @cindex Volume binding @cindex Unix namespace @cindex Namespace @cindex Binding names to filesystems Unix implements a namespace of hierarchically mounted filesystems. Two forms of binding between names and files are provided. A @dfn{hard link} completes the binding when the name is added to the filesystem. A @dfn{soft link} delays the binding until the name is accessed. An @dfn{automounter} adds a further form in which the binding of name to filesystem is delayed until the name is accessed.@refill The target volume, in its general form, is a tuple (host, filesystem, sublink) which can be used to name the physical location of any volume in the network. When a target is referenced, @i{Amd} ignores the sublink element and determines whether the required filesystem is already mounted. This is done by computing the local mount point for the filesystem and checking for an existing filesystem mounted at the same place. If such a filesystem already exists then it is assumed to be functionally identical to the target filesystem. By default there is a one-to-one mapping between the pair (host, filesystem) and the local mount point so this assumption is valid. @node Operational Principles, Mounting a Volume, Volume Binding, Overview @comment node-name, next, previous, up @section Operational Principles @cindex Operational principles @i{Amd} operates by introducing new mount points into the namespace. These are called @dfn{automount} points. The kernel sees these automount points as NFS filesystems being served by @i{Amd}. Having attached itself to the namespace, @i{Amd} is now able to control the view the rest of the system has of those mount points. RPC calls are received from the kernel one at a time. When a @dfn{lookup} call is received @i{Amd} checks whether the name is already known. If it is not, the required volume is mounted. A symbolic link pointing to the volume root is then returned. Once the symbolic link is returned, the kernel will send all other requests direct to the mounted filesystem. If a volume is not yet mounted, @i{Amd} consults a configuration @dfn{mount-map} corresponding to the automount point. @i{Amd} then makes a runtime decision on what and where to mount a filesystem based on the information obtained from the map. @i{Amd} does not implement all the NFS requests; only those relevant to name binding such as @dfn{lookup}, @dfn{readlink} and @dfn{readdir}. Some other calls are also implemented but most simply return an error code; for example @dfn{mkdir} always returns ``read-only filesystem''. @node Mounting a Volume, Automatic Unmounting, Operational Principles, Overview @comment node-name, next, previous, up @section Mounting a Volume @cindex Mounting a volume @cindex Location lists @cindex Alternate locations @cindex Mount retries @cindex Background mounts Each automount point has a corresponding mount map. The mount map contains a list of key--value pairs. The key is the name of the volume to be mounted. The value is a list of locations describing where the filesystem is stored in the network. In the source for the map the value would look like @display location1 location2 @dots{} locationN @end display @i{Amd} examines each location in turn. Each location may contain @dfn{selectors} which control whether @i{Amd} can use that location. For example, the location may be restricted to use by certain hosts. Those locations which cannot be used are ignored. @i{Amd} attempts to mount the filesystem described by each remaining location until a mount succeeds or @i{Amd} can no longer proceed. The latter can occur in three ways: @itemize @bullet @item If none of the locations could be used, or if all of the locations caused an error, then the last error is returned. @item If a location could be used but was being mounted in the background then @i{Amd} marks that mount as being ``in progress'' and continues with the next request; no reply is sent to the kernel. @item Lastly, one or more of the mounts may have been @dfn{deferred}. A mount is deferred if extra information is required before the mount can proceed. When the information becomes available the mount will take place, but in the mean time no reply is sent to the kernel. If the mount is deferred, @i{Amd} continues to try any remaining locations. @end itemize Once a volume has been mounted, @i{Amd} establishes a @dfn{volume mapping} which is used to satisfy subsequent requests.@refill @node Automatic Unmounting, Keep-alives, Mounting a Volume, Overview @comment node-name, next, previous, up @section Automatic Unmounting To avoid an ever increasing number of filesystem mounts, @i{Amd} removes volume mappings which have not been used recently. A time-to-live interval is associated with each mapping and when that expires the mapping is removed. When the last reference to a filesystem is removed, that filesystem is unmounted. If the unmount fails, for example the filesystem is still busy, the mapping is re-instated and its time-to-live interval is extended. The global default for this grace period is controlled by the ``-w'' command-line option (@pxref{-w Option, -w}). It is also possible to set this value on a per-mount basis (@pxref{opts Option, opts, opts}).@refill Filesystems can be forcefully timed out using the @i{Amq} command. @xref{Run-time Administration}. @node Keep-alives, Non-blocking Operation, Automatic Unmounting, Overview @comment node-name, next, previous, up @section Keep-alives @cindex Keep-alives @cindex Server crashes @cindex NFS ping Use of some filesystem types requires the presence of a server on another machine. If a machine crashes then it is of no concern to processes on that machine that the filesystem is unavailable. However, to processes on a remote host using that machine as a fileserver this event is important. This situation is most widely recognised when an NFS server crashes and the behaviour observed on client machines is that more and more processes hang. In order to provide the possibility of recovery, @i{Amd} implements a @dfn{keep-alive} interval timer for some filesystem types. Currently only NFS makes use of this service. The basis of the NFS keep-alive implementation is the observation that most sites maintain replicated copies of common system data such as manual pages, most or all programs, system source code and so on. If one of those servers goes down it would be reasonable to mount one of the others as a replacement. The first part of the process is to keep track of which fileservers are up and which are down. @i{Amd} does this by sending RPC requests to the servers' NFS @code{NullProc} and checking whether a reply is returned. While the server state is uncertain the requests are re-transmitted at three second intervals and if no reply is received after four attempts the server is marked down. If a reply is received the fileserver is marked up and stays in that state for 30 seconds at which time another NFS ping is sent. Once a fileserver is marked down, requests continue to be sent every 30 seconds in order to determine when the fileserver comes back up. During this time any reference through @i{Amd} to the filesystems on that server fail with the error ``Operation would block''. If a replacement volume is available then it will be mounted, otherwise the error is returned to the user. @c @i{Amd} keeps track of which servers are up and which are down. @c It does this by sending RPC requests to the servers' NFS {\sc NullProc} and @c checking whether a reply is returned. If no replies are received after a @c short period, @i{Amd} marks the fileserver @dfn{down}. @c RPC requests continue to be sent so that it will notice when a fileserver @c comes back up. @c ICMP echo packets \cite{rfc:icmp} are not used because it is the availability @c of the NFS service that is important, not the existence of a base kernel. @c Whenever a reference to a fileserver which is down is made via @i{Amd}, an alternate @c filesystem is mounted if one is available. @c Although this action does not protect user files, which are unique on the network, or processes which do not access files via @i{Amd} or already have open files on the hung filesystem, it can prevent most new processes from hanging. By default, fileserver state is not maintained for NFS/TCP mounts. The remote fileserver is always assumed to be up. @c @c With a suitable combination of filesystem management and mount-maps, @c machines can be protected against most server downtime. This can be @c enhanced by allocating boot-servers dynamically which allows a diskless @c workstation to be quickly restarted if necessary. Once the root filesystem @c is mounted, @i{Amd} can be started and allowed to mount the remainder of @c the filesystem from whichever fileservers are available. @node Non-blocking Operation, , Keep-alives, Overview @comment node-name, next, previous, up @section Non-blocking Operation @cindex Non-blocking operation @cindex Multiple-threaded server @cindex RPC retries Since there is only one instance of @i{Amd} for each automount point, and usually only one instance on each machine, it is important that it is always available to service kernel calls. @i{Amd} goes to great lengths to ensure that it does not block in a system call. As a last resort @i{Amd} will fork before it attempts a system call that may block indefinitely, such as mounting an NFS filesystem. Other tasks such as obtaining filehandle information for an NFS filesystem, are done using a purpose built non-blocking RPC library which is integrated with @i{Amd}'s task scheduler. This library is also used to implement NFS keep-alives (@pxref{Keep-alives}). Whenever a mount is deferred or backgrounded, @i{Amd} must wait for it to complete before replying to the kernel. However, this would cause @i{Amd} to block waiting for a reply to be constructed. Rather than do this, @i{Amd} simply @dfn{drops} the call under the assumption that the kernel RPC mechanism will automatically retry the request. @node Supported Platforms, Mount Maps, Overview, Top @comment node-name, next, previous, up @chapter Supported Platforms @i{Amd} has been ported to a wide variety of machines and operating systems. The table below lists those platforms supported by the current release. @menu * Supported Operating Systems:: * Supported Machine Architectures:: @end menu @node Supported Operating Systems, Supported Machine Architectures, Supported Platforms, Supported Platforms @comment node-name, next, previous, up @section Supported Operating Systems @cindex Operating system names @cindex Operating systems supported by Amd @cindex Supported operating systems The following operating systems are currently supported by @i{Amd}. @i{Amd}'s conventional name for each system is given. @table @code @item acis43 4.3 BSD for IBM RT. Contributed by Jan-Simon Pendry @t{} @item aix3 AIX 3.1. Contributed by Jan-Simon Pendry @t{} @item aux System V for Mac-II. Contributed by Julian Onions @t{} @item bsd44 4.4 BSD. Contributed by Jan-Simon Pendry @t{} @item concentrix Concentrix 5.0. Contributed by Sjoerd Mullender @t{} @item convex Convex OS 7.1. Contributed by Eitan Mizrotsky @t{} @item dgux Data General DG/UX. Contributed by Mark Davies @t{} @item fpx4 Celerity FPX 4.1/2. Contributed by Stephen Pope @t{} @item hcx Harris HCX/UX. Contributed by Chris Metcalf @t{} @item hlh42 HLH OTS 1.@i{x} (4.2 BSD). Contributed by Jan-Simon Pendry @t{} @item hpux HP-UX 6.@i{x} or 7.0. Contributed by Jan-Simon Pendry @t{} @item irix SGI Irix. Contributed by Scott R. Presnell @t{} @item next Mach for NeXT. Contributed by Bill Trost @t{} @item pyrOSx Pyramid OSx. Contributed by Stefan Petri @t{} @item riscix Acorn RISC iX. Contributed by Piete Brooks @t{} @item sos3 SunOS 3.4 & 3.5. Contributed by Jan-Simon Pendry @t{} @item sos4 SunOS 4.@i{x}. Contributed by Jan-Simon Pendry @t{} @item u2_2 Ultrix 2.2. Contributed by Piete Brooks @t{} @item u3_0 Ultrix 3. Contributed by Piete Brooks @t{} @item u4_0 Ultrix 4.0. Contributed by Chris Lindblad @t{} @item umax43 Umax 4.3 BSD. Contributed by Sjoerd Mullender @t{} @item utek Utek 4.0. Contributed by Bill Trost @t{} @item xinu43 mt Xinu MORE/bsd. Contributed by Jan-Simon Pendry @t{} @end table @node Supported Machine Architectures, , Supported Operating Systems, Supported Platforms @comment node-name, next, previous, up @section Supported Machine Architectures @cindex Supported machine architectures @cindex Machine architecture names @cindex Machine architectures supported by Amd @table @code @item alliant Alliant FX/4 @item arm Acorn ARM @item aviion Data General AViiON @item encore Encore @item fps500 FPS Model 500 @item hp9000 HP 9000/300 family @item hp9k8 HP 9000/800 family @item ibm032 IBM RT @item ibm6000 IBM RISC System/6000 @item iris4d SGI Iris 4D @item macII Apple Mac II @item mips MIPS RISC @item multimax Encore Multimax @item orion105 HLH Orion 1/05 @item sun3 Sun-3 family @item sun4 Sun-4 family @item tahoe Tahoe family @item vax DEC Vax @end table @node Mount Maps, Amd Command Line Options, Supported Platforms, Top @comment node-name, next, previous, up @chapter Mount Maps @cindex Mount maps @cindex Automounter configuration maps @cindex Mount information @i{Amd} has no built-in knowledge of machines or filesystems. External @dfn{mount-maps} are used to provide the required information. Specifically, @i{Amd} needs to know when and under what conditions it should mount filesystems. The map entry corresponding to the requested name contains a list of possible locations from which to resolve the request. Each location specifies filesystem type, information required by that filesystem (for example the block special device in the case of UFS), and some information describing where to mount the filesystem (@pxref{fs Option}). A location may also contain @dfn{selectors} (@pxref{Selectors}).@refill @menu * Map Types:: * Key Lookup:: * Location Format:: @end menu @node Map Types, Key Lookup, Mount Maps, Mount Maps @comment node-name, next, previous, up @section Map Types @cindex Mount map types @cindex Map types @cindex Configuration map types @cindex Types of mount map @cindex Types of configuration map @cindex Determining the map type A mount-map provides the run-time configuration information to @i{Amd}. Maps can be implemented in many ways. Some of the forms supported by @i{Amd} are regular files, NIS maps the @dfn{Hesiod} name server and even the password file. A mount-map @dfn{name} is a sequence of characters. When an automount point is created a handle on the mount-map is obtained. For each map type configured @i{Amd} attempts to reference the a map of the appropriate type. If a map is found, @i{Amd} notes the type for future use and deletes the reference, for example closing any open file descriptors. The available maps are configure when @i{Amd} is built and can be displayed by running the command @samp{amd -v}. By default, @i{Amd} caches data in a mode dependent on the type of map. This is the same as specifying @samp{cache:=mapdefault} and selects a suitable default cache mode depending on the map type. The individual defaults are described below. The @var{cache} option can be specified on automount points to alter the caching behaviour (@pxref{Automount Filesystem}).@refill The following map types have been implemented, though some are not available on all machines. Run the command @samp{amd -v} to obtain a list of map types configured on your machine. @menu * File maps:: * NIS maps:: * Hesiod maps:: * Password maps:: * Union maps:: @end menu @node File maps, NIS maps, Map Types, Map Types @comment node-name, next, previous, up @subsection File maps @cindex File maps @cindex Flat file maps @cindex File map syntactic conventions When @i{Amd} searches a file for a map entry it does a simple scan of the file and supports both comments and continuation lines. Continuation lines are indicated by a backslash character (@samp{\}) as the last character of a line in the file. The backslash, newline character @emph{and any leading white space on the following line} are discarded. A maximum line length of 2047 characters is enforced after continuation lines are read but before comments are stripped. Each line must end with a newline character; that is newlines are terminators, not separators. The following examples illustrate this: @example key valA valB; \ valC @end example specifies @emph{three} locations, and is identical to @example key valA valB; valC @end example However, @example key valA valB;\ valC @end example specifies only @emph{two} locations, and is identical to @example key valA valB;valC @end example After a complete line has been read from the file, including continuations, @i{Amd} determines whether there is a comment on the line. A comment begins with a hash (``@samp{#}'') character and continues to the end of the line. There is no way to escape or change the comment lead-in character. Note that continuation lines and comment support @dfn{only} apply to file maps. When caching is enabled, file maps have a default cache mode of @code{all} (@pxref{Automount Filesystem}). @node NIS maps, Hesiod maps, File maps, Map Types @comment node-name, next, previous, up @subsection NIS maps @cindex NIS (YP) maps When using NIS (formerly YP), an @i{Amd} map is implemented directly by the underlying NIS map. Comments and continuation lines are @emph{not} supported in the automounter and must be stripped when constructing the NIS server's database. NIS maps do not support cache mode @code{all} and, when caching is enabled, have a default cache mode of @code{inc} (@pxref{Automount Filesystem}). The following rule illustrates what could be added to your NIS @file{Makefile}, in this case causing the @file{amd.home} map to be rebuilt: @example $(YPTSDIR)/amd.home.time: $(ETCDIR)/amd.home -@@sed -e "s/#.*$$//" -e "/^$$/d" $(ETCDIR)/amd.home | \ awk '@{ \ for (i = 1; i <= NF; i++) \ if (i == NF) @{ \ if (substr($$i, length($$i), 1) == "\\") \ printf("%s", substr($$i, 1, length($$i) - 1)); \ else \ printf("%s\n", $$i); \ @} \ else \ printf("%s ", $$i); \ @}' | \ $(MAKEDBM) - $(YPDBDIR)/amd.home; \ touch $(YPTSDIR)/amd.home.time; \ echo "updated amd.home"; \ if [ ! $(NOPUSH) ]; then \ $(YPPUSH) amd.home; \ echo "pushed amd.home"; \ else \ : ; \ fi @end example Here @code{$(YPTSDIR)} contains the time stamp files, and @code{$(YPDBDIR)} contains the dbm format NIS files. @node Hesiod maps, Password maps, NIS maps, Map Types @comment node-name, next, previous, up @subsection Hesiod maps @cindex Hesiod maps When the map name begins with the string @samp{hesiod.} lookups are made using the @dfn{Hesiod} name server. The string following the dot is used as a name qualifier and is prepended with the key being located. The entire string is then resolved in the @code{automount} context. For example, if the key is @samp{jsp} and map name is @samp{hesiod.homes} then @dfn{Hesiod} is asked to resolve @samp{jsp.homes.automount}. Hesiod maps do not support cache mode @samp{all} and, when caching is enabled, have a default cache mode of @samp{inc} (@pxref{Automount Filesystem}). The following is an example of a @dfn{Hesiod} map entry: @example jsp.homes.automount HS TXT "rfs:=/home/charm;rhost:=charm;sublink:=jsp" njw.homes.automount HS TXT "rfs:=/home/dylan/dk2;rhost:=dylan;sublink:=njw" @end example @node Password maps, Union maps, Hesiod maps, Map Types @comment node-name, next, previous, up @subsection Password maps @cindex Password file maps @cindex /etc/passwd maps @cindex User maps, automatic generation @cindex Automatic generation of user maps @cindex Using the password file as a map The password map support is unlike the four previous map types. When the map name is the string @file{/etc/passwd} @i{Amd} can lookup a user name in the password file and re-arrange the home directory field to produce a usable map entry. @i{Amd} assumes the home directory has the format `@t{/}@i{anydir}@t{/}@i{dom1}@t{/../}@i{domN}@t{/}@i{login}'. @c @footnote{This interpretation is not necessarily exactly what you want.} It breaks this string into a map entry where @code{$@{rfs@}} has the value `@t{/}@i{anydir}@t{/}@i{domN}', @code{$@{rhost@}} has the value `@i{domN}@t{.}@i{...}@t{.}@i{dom1}', and @code{$@{sublink@}} has the value @samp{login}.@refill Thus if the password file entry was @example /home/achilles/jsp @end example the map entry used by @i{Amd} would be @example rfs:=/home/achilles;rhost:=achilles;sublink:=jsp @end example Similarly, if the password file entry was @example /home/cc/sugar/mjh @end example the map entry used by @i{Amd} would be @example rfs:=/home/sugar;rhost:=sugar.cc;sublink:=jsp @end example @node Union maps, , Password maps, Map Types @comment node-name, next, previous, up @subsection Union maps @cindex Union file maps The union map support is provided specifically for use with the union filesystem, @pxref{Union Filesystem}. It is identified by the string @samp{union:} which is followed by a colon separated list of directories. The directories are read in order, and the names of all entries are recorded in the map cache. Later directories take precedence over earlier ones. The union filesystem type then uses the map cache to determine the union of the names in all the directories. @c subsection Gdbm @node Key Lookup, Location Format, Map Types, Mount Maps @comment node-name, next, previous, up @section How keys are looked up @cindex Key lookup @cindex Map lookup @cindex Looking up keys @cindex How keys are looked up @cindex Wildcards in maps The key is located in the map whose type was determined when the automount point was first created. In general the key is a pathname component. In some circumstances this may be modified by variable expansion (@pxref{Variable Expansion}) and prefixing. If the automount point has a prefix, specified by the @var{pref} option, then that is prepended to the search key before the map is searched. If the map cache is a @samp{regexp} cache then the key is treated as an egrep-style regular expression, otherwise a normal string comparison is made. If the key cannot be found then a @dfn{wildcard} match is attempted. @i{Amd} repeatedly strips the basename from the key, appends @samp{/*} and attempts a lookup. Finally, @i{Amd} attempts to locate the special key @samp{*}. @group For example, the following sequence would be checked if @file{home/dylan/dk2} was being located: @example home/dylan/dk2 home/dylan/* home/* * @end example @end group At any point when a wildcard is found, @i{Amd} proceeds as if an exact match had been found and the value field is then used to resolve the mount request, otherwise an error code is propagated back to the kernel. (@pxref{Filesystem Types}).@refill @node Location Format, , Key Lookup, Mount Maps @comment node-name, next, previous, up @section Location Format @cindex Location format @cindex Map entry format @cindex How locations are parsed The value field from the lookup provides the information required to mount a filesystem. The information is parsed according to the syntax shown below. @display @i{location-list}: @i{location-selection} @i{location-list} @i{white-space} @t{||} @i{white-space} @i{location-selection} @i{location-selection}: @i{location} @i{location-selection} @i{white-space} @i{location} @i{location}: @i{location-info} @t{-}@i{location-info} @t{-} @i{location-info}: @i{sel-or-opt} @i{location-info}@t{;}@i{sel-or-opt} @t{;} @i{sel-or-opt}: @i{selection} @i{opt-ass} @i{selection}: selector@t{==}@i{value} selector@t{!=}@i{value} @i{opt-ass}: option@t{:=}@i{value} @i{white-space}: space tab @end display Note that unquoted whitespace is not allowed in a location description. White space is only allowed, and is mandatory, where shown with non-terminal @samp{white-space}. A @dfn{location-selection} is a list of possible volumes with which to satisfy the request. @dfn{location-selection}s are separated by the @samp{||} operator. The effect of this operator is to prevent use of location-selections to its right if any of the location-selections on its left were selected whether or not any of them were successfully mounted (@pxref{Selectors}).@refill The location-selection, and singleton @dfn{location-list}, @samp{type:=ufs;dev:=/dev/xd1g} would inform @i{Amd} to mount a UFS filesystem from the block special device @file{/dev/xd1g}. The @dfn{sel-or-opt} component is either the name of an option required by a specific filesystem, or it is the name of a built-in, predefined selector such as the architecture type. The value may be quoted with double quotes @samp{"}, for example @samp{type:="ufs";dev:="/dev/xd1g"}. These quotes are stripped when the value is parsed and there is no way to get a double quote into a value field. Double quotes are used to get white space into a value field, which is needed for the program filesystem (@pxref{Program Filesystem}).@refill @menu * Map Defaults:: * Variable Expansion:: * Selectors:: * Map Options:: @end menu @node Map Defaults, Variable Expansion, Location Format, Location Format @comment node-name, next, previous, up @subsection Map Defaults @cindex Map defaults @cindex How to set default map parameters @cindex Setting default map parameters A location beginning with a dash @samp{-} is used to specify default values for subsequent locations. Any previously specified defaults in the location-list are discarded. The default string can be empty in which case no defaults apply. The location @samp{-fs:=/mnt;opts:=ro} would set the local mount point to @file{/mnt} and cause mounts to be read-only by default. Defaults specified this way are appended to, and so override, any global map defaults given with @samp{/defaults}). @c @c A @samp{/defaults} value @dfn{gdef} and a location list @c \begin{quote} @c $@samp{-}@dfn{def}_a $\verb*+ +$ @dfn{loc}_{a_1} $\verb*+ +$ @dfn{loc}_{a_2} $\verb*+ +$ @samp{-}@dfn{def}_b $\verb*+ +$ @dfn{loc}_{b_1} \ldots$ @c \end{quote} @c is equivalent to @c \begin{quote} @c $@samp{-}@dfn{gdef}@samp{;}@dfn{def}_a $\verb*+ +$ @dfn{loc}_{a_1} $\verb*+ +$ @dfn{loc}_{a_2} $\verb*+ +$ @samp{-}@dfn{gdef}@samp{;}@dfn{def}_b $\verb*+ +$ @dfn{loc}_{b_1} \ldots$ @c \end{quote} @c which is equivalent to @c \begin{quote} @c $@dfn{gdef}@samp{;}@dfn{def}_a@samp{;}@dfn{loc}_{a_1} $\verb*+ +$@dfn{gdef}@samp{;}@dfn{def}_a@samp{;}@dfn{loc}_{a_2} $\verb*+ +$@dfn{gdef}@samp{;}@dfn{def}_b@samp{;}@dfn{loc}_{b_1} \ldots$ @c \end{quote} @node Variable Expansion, Selectors, Map Defaults, Location Format @comment node-name, next, previous, up @subsection Variable Expansion @cindex Variable expansion @cindex How variables are expanded @cindex Pathname operators @cindex Domain stripping @cindex Domainname operators @cindex Stripping the local domain name @cindex Environment variables @cindex How to access environment variables in maps To allow generic location specifications @i{Amd} does variable expansion on each location and also on some of the option strings. Any option or selector appearing in the form @code{$@dfn{var}} is replaced by the current value of that option or selector. For example, if the value of @code{$@{key@}} was @samp{bin}, @code{$@{autodir@}} was @samp{/a} and @code{$@{fs@}} was `@t{$@{autodir@}}@t{/local/}@t{$@{key@}}' then after expansion @code{$@{fs@}} would have the value @samp{/a/local/bin}. Any environment variable can be accessed in a similar way.@refill Two pathname operators are available when expanding a variable. If the variable name begins with @samp{/} then only the last component of then pathname is substituted. For example, if @code{$@{path@}} was @samp{/foo/bar} then @code{$@{/path@}} would be expanded to @samp{bar}. Similarly, if the variable name ends with @samp{/} then all but the last component of the pathname is substituted. In the previous example, @code{$@{path/@}} would be expanded to @samp{/foo}.@refill Two domain name operators are also provided. If the variable name begins with @samp{.} then only the domain part of the name is substituted. For example, if @code{$@{rhost@}} was @samp{swan.doc.ic.ac.uk} then @code{$@{.rhost@}} would be expanded to @samp{doc.ic.ac.uk}. Similarly, if the variable name ends with @samp{.} then only the host component is substituted. In the previous example, @code{$@{rhost.@}} would be expanded to @samp{swan}.@refill Variable expansion is a two phase process. Before a location is parsed, all references to selectors, @i{eg} @code{$@{path@}}, are expanded. The location is then parsed, selections are evaluated and option assignments recorded. If there were no selections or they all succeeded the location is used and the values of the following options are expanded in the order given: @var{sublink}, @var{rfs}, @var{fs}, @var{opts}, @var{remopts}, @var{mount} and @var{unmount}. Note that expansion of option values is done after @dfn{all} assignments have been completed and not in a purely left to right order as is done by the shell. This generally has the desired effect but care must be taken if one of the options references another, in which case the ordering can become significant. There are two special cases concerning variable expansion: @enumerate @item before a map is consulted, any selectors in the name received from the kernel are expanded. For example, if the request from the kernel was for `@t{$@{arch@}}@t{.bin}' and the machine architecture was @samp{vax}, the value given to @code{$@{key@}} would be @samp{vax.bin}.@refill @item the value of @code{$@{rhost@}} is expanded and normalized before the other options are expanded. The normalization process strips any local sub-domain components. For example, if @code{$@{domain@}} was @samp{Berkeley.EDU} and @code{$@{rhost@}} was initially @samp{snow.Berkeley.EDU}, after the normalization it would simply be @samp{snow}. Hostname normalization is currently done in a @emph{case-dependent} manner.@refill @end enumerate @node Selectors, Map Options, Variable Expansion, Location Format @comment node-name, next, previous, up @subsection Selectors @cindex Selectors Selectors are used to control the use of a location. It is possible to share a mount map between many machines in such a way that filesystem location, architecture and operating system differences are hidden from the users. A selector of the form @samp{arch==sun3;os==sos4} would only apply on Sun-3s running SunOS 4.x. Selectors are evaluated left to right. If a selector fails then that location is ignored. Thus the selectors form a conjunction and the locations form a disjunction. If all the locations are ignored or otherwise fail then @i{Amd} uses the @dfn{error} filesystem (@pxref{Error Filesystem}). This is equivalent to having a location @samp{type:=error} at the end of each mount-map entry.@refill The selectors currently implemented are: @table @samp @cindex arch, mount selector @cindex Mount selector; arch @cindex Selector; arch @item arch the machine architecture which was automatically determined at compile time. The architecture type can be displayed by running the command @samp{amd -v}. @xref{Supported Machine Architectures}.@refill @item autodir @cindex autodir, mount selector @cindex Mount selector; autodir @cindex Selector; autodir the default directory under which to mount filesystems. This may be changed by the ``-a'' command line option. See the @var{fs} option. @item byte @cindex byte, mount selector @cindex Mount selector; byte @cindex Selector; byte the machine's byte ordering. This is either @samp{little}, indicating little-endian, or @samp{big}, indicating big-endian. One possible use is to share @samp{rwho} databases (@pxref{rwho servers}). @item cluster @cindex cluster, mount selector @cindex Mount selector; cluster @cindex Selector; cluster is provided as a hook for the name of the local cluster. This can be used to decide which servers to use for copies of replicated filesystems. @code{$@{cluster@}} defaults to the value of @code{$@{domain@}} unless a different value is set with the ``-C'' command line option. @item domain @cindex domain, mount selector @cindex Mount selector; domain @cindex Selector; domain the local domain name as specified by the ``-d'' command line option. See @samp{host}. @item host @cindex host, mount selector @cindex Mount selector; host @cindex Selector; host the local hostname as determined by @b{gethostname}(2). If no domain name was specified on the command line and the hostname contains a period @samp{.} then the string before the period is used as the host name, and the string after the period is assigned to @code{$@{domain@}}. For example, if the hostname is @samp{styx.doc.ic.ac.uk} then @code{host} would be @samp{styx} and @code{domain} would be @samp{doc.ic.ac.uk}. @code{hostd} would be @samp{styx.doc.ic.ac.uk}.@refill @item hostd @cindex hostd, mount selector @cindex Mount selector; hostd @cindex Selector; hostd is @code{$@{host@}} and @code{$@{domain@}} concatenated with a @samp{.} inserted between them if required. If @code{$@{domain@}} is an empty string then @code{$@{host@}} and @code{$@{hostd@}} will be identical. @item karch @cindex karch, mount selector @cindex Mount selector; karch @cindex Selector; karch is provided as a hook for the kernel architecture. This is used on SunOS 4, for example, to distinguish between different @samp{/usr/kvm} volumes. @code{$@{karch@}} defaults to the value of @code{$@{arch@}} unless a different value is set with the ``-k'' command line option. @item os @cindex os, mount selector @cindex Mount selector; os @cindex Selector; os the operating system. Like the machine architecture, this is automatically determined at compile time. The operating system name can be displayed by running the command @samp{amd -v}. @xref{Supported Operating Systems}.@refill @end table The following selectors are also provided. Unlike the other selectors, they vary for each lookup. Note that when the name from the kernel is expanded prior to a map lookup, these selectors are all defined as empty strings. @table @samp @item key @cindex key, mount selector @cindex Mount selector; key @cindex Selector; key the name being resolved. For example, if @file{/home} is an automount point, then accessing @file{/home/foo} would set @code{$@{key@}} to the string @samp{foo}. The key is prefixed by the @var{pref} option set in the parent mount point. The default prefix is an empty string. If the prefix was @file{blah/} then @code{$@{key@}} would be set to @file{blah/foo}.@refill @item map @cindex map, mount selector @cindex Mount selector; map @cindex Selector; map the name of the mount map being used. @item path @cindex path, mount selector @cindex Mount selector; path @cindex Selector; path the full pathname of the name being resolved. For example @file{/home/foo} in the example above. @item wire @cindex wire, mount selector @cindex Mount selector; wire @cindex Selector; wire the name of the network to which the primary network interface is attached. If a symbolic name cannot be found in the networks or hosts database then dotted IP address format is used. This value is also output by the ``-v'' option. @end table Selectors can be negated by using @samp{!=} instead of @samp{==}. For example to select a location on all non-Vax machines the selector @samp{arch!=vax} would be used. @node Map Options, , Selectors, Location Format @comment node-name, next, previous, up @subsection Map Options @cindex Map options @cindex Setting map options Options are parsed concurrently with selectors. The difference is that when an option is seen the string following the @samp{:=} is recorded for later use. As a minimum the @var{type} option must be specified. Each filesystem type has other options which must also be specified. @xref{Filesystem Types}, for details on the filesystem specific options.@refill Superfluous option specifications are ignored and are not reported as errors. The following options apply to more than one filesystem type. @menu * delay Option:: * fs Option:: * opts Option:: * remopts Option:: * sublink Option:: * type Option:: @end menu @node delay Option, fs Option, Map Options, Map Options @comment node-name, next, previous, up @subsubsection delay Option @cindex Setting a delay on a mount location @cindex Delaying mounts from specific locations @cindex Primary server @cindex Secondary server @cindex delay, mount option @cindex Mount option; delay The delay, in seconds, before an attempt will be made to mount from the current location. Auxilliary data, such as network address, file handles and so on are computed regardless of this value. A delay can be used to implement the notion of primary and secondary file servers. The secondary servers would have a delay of a few seconds, thus giving the primary servers a chance to respond first. @node fs Option, opts Option, delay Option, Map Options @comment node-name, next, previous, up @subsubsection fs Option @cindex Setting the local mount point @cindex Overriding the default mount point @cindex fs, mount option @cindex Mount option; fs The local mount point. The semantics of this option vary between filesystems. For NFS and UFS filesystems the value of @code{$@{fs@}} is used as the local mount point. For other filesystem types it has other meanings which are described in the section describing the respective filesystem type. It is important that this string uniquely identifies the filesystem being mounted. To satisfy this requirement, it should contain the name of the host on which the filesystem is resident and the pathname of the filesystem on the local or remote host. The reason for requiring the hostname is clear if replicated filesystems are considered. If a fileserver goes down and a replacement filesystem is mounted then the @dfn{local} mount point @dfn{must} be different from that of the filesystem which is hung. Some encoding of the filesystem name is required if more than one filesystem is to be mounted from any given host. If the hostname is first in the path then all mounts from a particular host will be gathered below a single directory. If that server goes down then the hung mount points are less likely to be accidentally referenced, for example when @b{getwd}(3) traverses the namespace to find the pathname of the current directory. The @samp{fs} option defaults to @code{$@{autodir@}/$@{rhost@}$@{rfs@}}. In addition, @samp{rhost} defaults to the local host name (@code{$@{host@}}) and @samp{rfs} defaults to the value of @code{$@{path@}}, which is the full path of the requested file; @samp{/home/foo} in the example above (@pxref{Selectors}). @code{$@{autodir@}} defaults to @samp{/a} but may be changed with the ``-a'' command line option. Sun's automounter defaults to @samp{/tmp_mnt}. Note that there is no @samp{/} between the @code{$@{rhost@}} and @code{$@{rfs@}} since @code{$@{rfs@}} begins with a @samp{/}.@refill @node opts Option, remopts Option, fs Option, Map Options @comment node-name, next, previous, up @subsubsection opts Option @cindex Setting system mount options @cindex Passing parameters to the mount system call @cindex mount system call @cindex mount system call flags @cindex The mount system call @cindex opts, mount option @cindex Mount option; opts The options to pass to the mount system call. A leading @samp{-} is silently ignored. The mount options supported generally correspond to those used by @b{mount}(8) and are listed below. Some additional pseudo-options are interpreted by @i{Amd} and are also listed. Unless specifically overridden, each of the system default mount options applies. Any options not recognised are ignored. If no options list is supplied the string @samp{rw,defaults} is used and all the system default mount options apply. Options which are not applicable for a particular operating system are silently ignored. For example, only 4.4 BSD is known to implement the @code{compress} and @code{spongy} options. @table @code @item compress Use NFS compression protocol. @item grpid Use BSD directory group-id semantics. @item intr Allow keyboard interrupts on hard mounts. @item noconn Don't make a connection on datagram transports. @item nocto No close-to-open consistency. @item nodevs Don't allow local special devices on this filesystem. @item nolooklease When nqnfs is selected, disable the automatic acquiring of leases on lookups. @item nosuid Don't allow set-uid or set-gid executables on this filesystem. @item nqnfs Enable leasing extensions to the NFS protocol know as Not Quite NFS. Only supported by the 4.4BSD NFS implementation. @item quota Enable quota checking on this mount. @item resvport Use a reserved port number (one less than 1024). Some NFS servers require this. @item retrans=@i{n} The number of NFS retransmits made before a user error is generated by a @samp{soft} mounted filesystem, and before a @samp{hard} mounted filesystem reports @samp{NFS server @dfn{yoyo} not responding still trying}. @item ro Mount this filesystem readonly. @item rsize=@var{n} The NFS read packet size. You may need to set this if you are using NFS/UDP through a gateway. @item soft Give up after @dfn{retrans} retransmissions. @item spongy Like @samp{soft} for status requests, and @samp{hard} for data transfers. @item tcp Use TCP/IP instead of UDP/IP, ignored if the NFS implementation does not support TCP/IP mounts. @item timeo=@var{n} The NFS timeout, in tenth-seconds, before a request is retransmitted. @item wsize=@var{n} The NFS write packet size. You may need to set this if you are using NFS/UDP through a gateway. @end table The following options are implemented by @i{Amd}, rather than being passed to the kernel. @table @code @item nounmount Configures the mount so that its time-to-live will never expire. This is also the default for some filesystem types. @c @c Implementation broken: @item ping=@var{n} The interval, in seconds, between keep-alive pings. When four consecutive pings have failed the mount point is marked as hung. This interval defaults to 30 seconds. If the ping interval is less than zero, no pings are sent and the host is assumed to be always up. By default, pings are not sent for an NFS/TCP mount. @item retry=@var{n} The number of times to retry the mount system call. @item utimeout=@var{n} The interval, in seconds, by which the mount's time-to-live is extended after an unmount attempt has failed. In fact the interval is extended before the unmount is attempted to avoid thrashing. The default value is 120 seconds (two minutes) or as set by the ``-w'' command line option. @end table @node remopts Option, sublink Option, opts Option, Map Options @comment node-name, next, previous, up @subsubsection remopts Option @cindex Setting system mount options for non-local networks @cindex remopts, mount option @cindex Mount option; remopts This option has the same use as @code{$@{opts@}} but applies only when the remote host is on a non-local network. For example, when using NFS across a gateway it is often necessary to use smaller values for the data read and write sizes. This can simply be done by specifying the small values in @var{remopts}. When a non-local host is accessed, the smaller sizes will automatically be used. @i{Amd} determines whether a host is local by examining the network interface configuration at startup. Any interface changes made after @i{Amd} has been started will not be noticed. The likely effect will be that a host may incorrectly be declared non-local. Unless otherwise set, the value of @code{$@{rem@}} is the same as the value of @code{$@{opts@}}. @node sublink Option, type Option, remopts Option, Map Options @comment node-name, next, previous, up @subsubsection sublink Option @cindex Setting the sublink option @cindex sublink, mount option @cindex Mount option; sublink The subdirectory within the mounted filesystem to which the reference should point. This can be used to prevent duplicate mounts in cases where multiple directories in the same mounted filesystem are used. @node type Option, , sublink Option, Map Options @comment node-name, next, previous, up @subsubsection type Option @cindex Setting the filesystem type option @cindex type, mount option @cindex Mount option; type The filesystem type to be used. @xref{Filesystem Types}, for a full description of each type.@refill @node Amd Command Line Options, Filesystem Types, Mount Maps, Top @comment node-name, next, previous, up @chapter @i{Amd} Command Line Options @cindex Command line options, Amd @cindex Amd command line options @cindex Overriding defaults on the command line Many of @i{Amd}'s parameters can be set from the command line. The command line is also used to specify automount points and maps. The general format of a command line is @example amd [@i{options}] @{ @i{directory} @i{map-name} [-@i{map-options}] @} ... @end example For each directory and map-name given, @i{Amd} establishes an automount point. The @dfn{map-options} may be any sequence of options or selectors---@pxref{Location Format}. The @dfn{map-options} apply only to @i{Amd}'s mount point. @samp{type:=toplvl;cache:=mapdefault;fs:=$@{map@}} is the default value for the map options. Default options for a map are read from a special entry in the map whose key is the string @samp{/defaults}. When default options are given they are prepended to any options specified in the mount-map locations as explained in. @xref{Map Defaults}, for more details. The @dfn{options} are any combination of those listed below. Once the command line has been parsed, the automount points are mounted. The mount points are created if they do not already exist, in which case they will be removed when @i{Amd} exits. Finally, @i{Amd} disassociates itself from its controlling terminal and forks into the background. Note: Even if @i{Amd} has been built with @samp{-DDEBUG} it will still background itself and disassociate itself from the controlling terminal. To use a debugger it is necessary to specify @samp{-D nodaemon} on the command line. @menu * -a Option:: Automount directory. * -c Option:: Cache timeout interval. * -d Option:: Domain name. * -k Option:: Kernel architecture. * -l Option:: Log file. * -n Option:: Hostname normalisation. * -p Option:: Output process id. * -r Option:: Restart existing mounts. * -t Option:: Kernel RPC timeout. * -v Option:: Version information. * -w Option:: Wait interval after failed unmount. * -x Option:: Log options. * -y Option:: NIS domain. * -C-Option:: Cluster name. * -D-Option:: Debug flags. @end menu @node -a Option, -c Option, Amd Command Line Options, Amd Command Line Options @comment node-name, next, previous, up @section @code{-a} @var{directory} @cindex Automount directory @cindex Setting the default mount directory Specifies the default mount directory. This option changes the variable @code{$@{autodir@}} which otherwise defaults to @file{/a}. For example, some sites prefer @file{/amd}. @example amd -a /amd ... @end example @node -c Option, -d Option, -a Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-c} @var{cache-interval} @cindex Cache interval @cindex Interval before a filesystem times out @cindex Setting the interval before a filesystem times out @cindex Changing the interval before a filesystem times out Selects the period, in seconds, for which a name is cached by @i{Amd}. If no reference is made to the volume in this period, @i{Amd} discards the volume name to filesystem mapping. Once the last reference to a filesystem has been removed, @i{Amd} attempts to unmount the filesystem. If the unmount fails the interval is extended by a further period as specified by the @samp{-w} command line option or by the @samp{utimeout} mount option. The default @dfn{cache-interval} is 300 seconds (five minutes). @node -d Option, -k Option, -c Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-d} @var{domain} @cindex Domain name @cindex Setting the local domain name @cindex Overriding the local domain name Specifies the host's domain. This sets the internal variable @code{$@{domain@}} and affects the @code{$@{hostd@}} variable. If this option is not specified and the hostname already contains the local domain then that is used, otherwise the default value of @code{$@{domain@}} is @samp{unknown.domain}. For example, if the local domain was @samp{doc.ic.ac.uk}, @i{Amd} could be started as follows: @example amd -d doc.ic.ac.uk ... @end example @node -k Option, -l Option, -d Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-k} @var{kernel-architecture} @cindex Setting the Kernel architecture Specifies the kernel architecture of the system. This is usually the output of @samp{arch -k} and its only effect is to set the variable @code{$@{karch@}}. If this option is not given, @code{$@{karch@}} has the same value as @code{$@{arch@}}. This would be used as follows: @example amd -k `arch -k` ... @end example @node -l Option, -n Option, -k Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-l} @var{log-option} @cindex Log filename @cindex Setting the log file @cindex Using syslog to log errors @cindex syslog Selects the form of logging to be made. Two special @dfn{log-options} are recognised. @enumerate @item If @dfn{log-option} is the string @samp{syslog}, @i{Amd} will use the @b{syslog}(3) mechanism.@refill @item If @dfn{log-option} is the string @samp{/dev/stderr}, @i{Amd} will use standard error, which is also the default target for log messages. To implement this, @i{Amd} simulates the effect of the @samp{/dev/fd} driver. @end enumerate Any other string is taken as a filename to use for logging. Log messages are appended to the file if it already exists, otherwise a new file is created. The file is opened once and then held open, rather than being re-opened for each message. If the @samp{syslog} option is specified but the system does not support syslog or if the named file cannot be opened or created, @i{Amd} will use standard error. Error messages generated before @i{Amd} has finished parsing the command line are printed on standard error. Using @samp{syslog} is usually best, in which case @i{Amd} would be started as follows: @example amd -l syslog ... @end example @node -n Option, -p Option, -l Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-n} @cindex Hostname normalisation @cindex Aliased hostnames @cindex Resolving aliased hostnames @cindex Normalising hostnames Normalises the remote hostname before using it. Normalisation is done by replacing the value of @code{$@{rhost@}} with the primary name returned by a hostname lookup. This option should be used if several names are used to refer to a single host in a mount map. @node -p Option, -r Option, -n Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-p} @cindex Process id @cindex Displaying the process id @cindex process id of Amd daemon @cindex pid file, creating with -p option @cindex Creating a pid file Causes @i{Amd}'s process id to be printed on standard output. This can be redirected to a suitable file for use with kill: @example amd -p > /var/run/amd.pid ... @end example This option only has an affect if @i{Amd} is running in daemon mode. If @i{Amd} is started with the @code{-D nodaemon} debug flag, this option is ignored. @node -r Option, -t Option, -p Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-r} @cindex Restarting existing mounts @cindex Picking up existing mounts Tells @i{Amd} to restart existing mounts (@pxref{Inheritance Filesystem}). @c @dfn{This option will be made the default in the next release.} @node -t Option, -v Option, -r Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-t} @var{timeout.retransmit} @cindex Setting Amd's RPC parameters Specifies the RPC @dfn{timeout} and @dfn{retransmit} intervals used by the kernel to communicate to @i{Amd}. These are used to set the @samp{timeo} and @samp{retrans} mount options. @i{Amd} relies on the kernel RPC retransmit mechanism to trigger mount retries. The value of this parameter changes the retry interval. Too long an interval gives poor interactive response, too short an interval causes excessive retries. @node -v Option, -w Option, -t Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-v} @cindex Version information @cindex Discovering version information @cindex How to discover your version of Amd Print version information on standard error and then exit. The output is of the form: @example amd 5.2.1.11 of 91/03/17 18:04:05 5.3Alpha11 #0: Sun Mar 17 18:07:28 GMT 1991 Built by pendry@@vangogh.Berkeley.EDU for a hp300 running bsd44 (big-endian). Map support for: root, passwd, union, file, error. FS: ufs, nfs, nfsx, host, link, program, union, auto, direct, toplvl, error. Primary network is 128.32.130.0. @end example The information includes the version number, release date and name of the release. The architecture (@pxref{Supported Machine Architectures}), operating system (@pxref{Supported Operating Systems}) and byte ordering are also printed as they appear in the @code{$@{os@}}, @code{$@{arch@}} and @code{$@{byte@}} variables.@refill @node -w Option, -x Option, -v Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-w} @var{wait-timeout} @cindex Setting the interval between unmount attempts @cindex unmount attempt backoff interval Selects the interval in seconds between unmount attempts after the initial time-to-live has expired. This defaults to 120 seconds (two minutes). @node -x Option, -y Option, -w Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-x} @var{opts} @cindex Log message selection @cindex Selecting specific log messages @cindex How to select log messages @cindex syslog priorities Specifies the type and verbosity of log messages. @dfn{opts} is a comma separated list selected from the following options: @table @code @item fatal Fatal errors @item error Non-fatal errors @item user Non-fatal user errors @item warn Recoverable errors @item warning Alias for @code{warn} @item info Information messages @item map Mount map usage @item stats Additional statistics @item all All of the above @end table Initially a set of default logging flags is enabled. This is as if @samp{-x all,nomap,nostats} had been selected. The command line is parsed and logging is controlled by the ``-x'' option. The very first set of logging flags is saved and can not be subsequently disabled using @i{Amq}. This default set of options is useful for general production use.@refill The @samp{info} messages include details of what is mounted and unmounted and when filesystems have timed out. If you want to have the default set of messages without the @samp{info} messages then you simply need @samp{-x noinfo}. The messages given by @samp{user} relate to errors in the mount maps, so these are useful when new maps are installed. The following table lists the syslog priorites used for each of the message types.@refill @table @code @item fatal LOG_CRIT @item error LOG_ERR @item user LOG_WARNING @item warning LOG_WARNING @item info LOG_INFO @item debug LOG_DEBUG @item map LOG_DEBUG @item stats LOG_INFO @end table The options can be prefixed by the string @samp{no} to indicate that this option should be turned off. For example, to obtain all but @samp{info} messages the option @samp{-x all,noinfo} would be used. If @i{Amd} was built with debugging enabled the @code{debug} option is automatically enabled regardless of the command line options. @node -y Option, -C-Option, -x Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-y} @var{NIS-domain} @cindex NIS (YP) domain name @cindex Overriding the NIS (YP) domain name @cindex Setting the NIS (YP) domain name @cindex YP domain name Selects an alternate NIS domain. This is useful for debugging and cross-domain shared mounting. If this flag is specified, @i{Amd} immediately attempts to bind to a server for this domain. @c @i{Amd} refers to NIS maps when it starts, unless the ``-m'' option @c is specified, and whenever required in a mount map. @node -C-Option, -D-Option, -y Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-C} @var{cluster-name} @cindex Cluster names @cindex Setting the cluster name Specifies the name of the cluster of which the local machine is a member. The only effect is to set the variable @code{$@{cluster@}}. The @dfn{cluster-name} is will usually obtained by running another command which uses a database to map the local hostname into a cluster name. @code{$@{cluster@}} can then be used as a selector to restrict mounting of replicated data. If this option is not given, @code{$@{cluster@}} has the same value as @code{$@{domain@}}. This would be used as follows: @example amd -C `clustername` ... @end example @node -D-Option, , -C-Option, Amd Command Line Options @comment node-name, next, previous, up @section @code{-D} @var{opts} @cindex Debug options @cindex Setting debug flags Controls the verbosity and coverage of the debugging trace; @dfn{opts} is a comma separated list of debugging options. The ``-D'' option is only available if @i{Amd} was compiled with @samp{-DDEBUG}. The memory debugging facilities are only available if @i{Amd} was compiled with @samp{-DDEBUG_MEM} (in addition to @samp{-DDEBUG}). The most common options to use are @samp{-D trace} and @samp{-D test} (which turns on all the useful debug options). See the program source for a more detailed explanation of the available options. @node Filesystem Types, Run-time Administration, Amd Command Line Options, Top @comment node-name, next, previous, up @chapter Filesystem Types @cindex Filesystem types @cindex Mount types @cindex Types of filesystem To mount a volume, @i{Amd} must be told the type of filesystem to be used. Each filesystem type typically requires additional information such as the fileserver name for NFS. From the point of view of @i{Amd}, a @dfn{filesystem} is anything that can resolve an incoming name lookup. An important feature is support for multiple filesystem types. Some of these filesystems are implemented in the local kernel and some on remote fileservers, whilst the others are implemented internally by @i{Amd}.@refill The two common filesystem types are UFS and NFS. Four other user accessible filesystems (@samp{link}, @samp{program}, @samp{auto} and @samp{direct}) are also implemented internally by @i{Amd} and these are described below. There are two additional filesystem types internal to @i{Amd} which are not directly accessible to the user (@samp{inherit} and @samp{error}). Their use is described since they may still have an effect visible to the user.@refill @menu * Network Filesystem:: A single NFS filesystem. * Network Host Filesystem:: NFS mount a host's entire export tree. * Network Filesystem Group:: An atomic group of NFS filesystems. * Unix Filesystem:: Native disk filesystem. * Program Filesystem:: Generic Program mounts. * Symbolic Link Filesystem:: Local link referencing existing filesystem. * Automount Filesystem:: * Direct Automount Filesystem:: * Union Filesystem:: * Error Filesystem:: * Top-level Filesystem:: * Root Filesystem:: * Inheritance Filesystem:: @end menu @node Network Filesystem, Network Host Filesystem, Filesystem Types, Filesystem Types @comment node-name, next, previous, up @section Network Filesystem (@samp{type:=nfs}) @cindex NFS @cindex Mounting an NFS filesystem @cindex How to mount and NFS filesystem @cindex nfs, filesystem type @cindex Filesystem type; nfs The @dfn{nfs} filesystem type provides access to Sun's NFS. @noindent The following options must be specified: @table @code @cindex rhost, mount option @cindex Mount option; rhost @item rhost the remote fileserver. This must be an entry in the hosts database. IP addresses are not accepted. The default value is taken from the local host name (@code{$@{host@}}) if no other value is specified. @cindex rfs, mount option @cindex Mount option; rfs @item rfs the remote filesystem. If no value is specified for this option, an internal default of @code{$@{path@}} is used. @end table NFS mounts require a two stage process. First, the @dfn{file handle} of the remote file system must be obtained from the server. Then a mount system call must be done on the local system. @i{Amd} keeps a cache of file handles for remote file systems. The cache entries have a lifetime of a few minutes. If a required file handle is not in the cache, @i{Amd} sends a request to the remote server to obtain it. @i{Amd} @dfn{does not} wait for a response; it notes that one of the locations needs retrying, but continues with any remaining locations. When the file handle becomes available, and assuming none of the other locations was successfully mounted, @i{Amd} will retry the mount. This mechanism allows several NFS filesystems to be mounted in parallel. @c @footnote{The mechanism @c is general, however NFS is the only filesystem @c for which the required hooks have been written.} The first one which responds with a valid file handle will be used. @noindent An NFS entry might be: @example jsp host!=charm;type:=nfs;rhost:=charm;rfs:=/home/charm;sublink:=jsp @end example The mount system call and any unmount attempts are always done in a new task to avoid the possibilty of blocking @i{Amd}. @node Network Host Filesystem, Network Filesystem Group, Network Filesystem, Filesystem Types @comment node-name, next, previous, up @section Network Host Filesystem (@samp{type:=host}) @cindex Network host filesystem @cindex Mounting entire export trees @cindex How to mount all NFS exported filesystems @cindex host, filesystem type @cindex Filesystem type; host @c NOTE: the current implementation of the @dfn{host} filesystem type @c sometimes fails to maintain a consistent view of the remote mount tree. @c This happens when the mount times out and only some of the remote mounts @c are successfully unmounted. To prevent this from occurring, use the @c @samp{nounmount} mount option. The @dfn{host} filesystem allows access to the entire export tree of an NFS server. The implementation is layered above the @samp{nfs} implementation so keep-alives work in the same way. The only option which needs to specified is @samp{rhost} which is the name of the fileserver to mount. The @samp{host} filesystem type works by querying the mount daemon on the given fileserver to obtain its export list. @i{Amd} then obtains filehandles for each of the exported filesystems. Any errors at this stage cause that particular filesystem to be ignored. Finally each filesystem is mounted. Again, errors are logged but ignored. One common reason for mounts to fail is that the mount point does not exist. Although @i{Amd} attempts to automatically create the mount point, it may be on a remote filesystem to which @i{Amd} does not have write permission. When an attempt to unmount a @samp{host} filesystem mount fails, @i{Amd} remounts any filesystems which had successfully been unmounted. To do this @i{Amd} queries the mount daemon again and obtains a fresh copy of the export list. @i{Amd} then tries to mount any exported filesystems which are not currently mounted. Sun's automounter provides a special @samp{-hosts} map. To achieve the same effect with @i{Amd} requires two steps. First a mount map must be created as follows: @example /defaults type:=host;fs:=$@{autodir@}/$@{rhost@}/root;rhost:=$@{key@} * opts:=rw,nosuid,grpid @end example @noindent and then start @i{Amd} with the following command @example amd /n net.map @end example @noindent where @samp{net.map} is the name of map described above. Note that the value of @code{$@{fs@}} is overridden in the map. This is done to avoid a clash between the mount tree and any other filesystem already mounted from the same fileserver. If different mount options are needed for different hosts then additional entries can be added to the map, for example @example host2 opts:=ro,nosuid,soft @end example @noindent would soft mount @samp{host2} read-only. @node Network Filesystem Group, Unix Filesystem, Network Host Filesystem, Filesystem Types @comment node-name, next, previous, up @section Network Filesystem Group (@samp{type:=nfsx}) @cindex Network filesystem group @cindex Atomic NFS mounts @cindex Mounting an atomic group of NFS filesystems @cindex How to mount an atomic group of NFS filesystems @cindex nfsx, filesystem type @cindex Filesystem type; nfsx The @dfn{nfsx} filesystem allows a group of filesystems to be mounted from a single NFS server. The implementation is layered above the @samp{nfs} implementation so keep-alives work in the same way. The options are the same as for the @samp{nfs} filesystem with one difference. @noindent The following options must be specified: @table @code @item rhost the remote fileserver. This must be an entry in the hosts database. IP addresses are not accepted. The default value is taken from the local host name (@code{$@{host@}}) if no other value is specified. @item rfs as a list of filesystems to mount. The list is in the form of a comma separated strings. @end table @noindent For example: @example pub type:=nfsx;rhost:=gould;\ rfs:=/public,/,graphics,usenet;fs:=$@{autodir@}/$@{rhost@}/root @end example The first string defines the root of the tree, and is applied as a prefix to the remaining members of the list which define the individual filesystems. The first string is @emph{not} used as a filesystem name. A parallel operation is used to determine the local mount points to ensure a consistent layout of a tree of mounts. Here, the @emph{three} filesystems, @samp{/public}, @samp{/public/graphics} and @samp{/public/usenet}, would be mounted.@refill A local mount point, @code{$@{fs@}}, @emph{must} be specified. The default local mount point will not work correctly in the general case. A suggestion is to use @samp{fs:=$@{autodir@}/$@{rhost@}/root}.@refill @node Unix Filesystem, Program Filesystem, Network Filesystem Group, Filesystem Types @comment node-name, next, previous, up @section Unix Filesystem (@samp{type:=ufs}) @cindex Unix filesystem @cindex UFS @cindex Mounting a UFS filesystem @cindex Mounting a local disk @cindex How to mount a UFS filesystems @cindex How to mount a local disk @cindex Disk filesystems @cindex ufs, filesystem type @cindex Filesystem type; ufs The @dfn{ufs} filesystem type provides access to the system's standard disk filesystem---usually a derivative of the Berkeley Fast Filesystem. @noindent The following option must be specified: @table @code @cindex dev, mount option @cindex Mount option; dev @item dev the block special device to be mounted. @end table A UFS entry might be: @example jsp host==charm;type:=ufs;dev:=/dev/xd0g;sublink:=jsp @end example @node Program Filesystem, Symbolic Link Filesystem, Unix Filesystem, Filesystem Types @comment node-name, next, previous, up @section Program Filesystem (@samp{type:=program}) @cindex Program filesystem @cindex Mount a filesystem under program control @cindex program, filesystem type @cindex Filesystem type; program The @dfn{program} filesystem type allows a program to be run whenever a mount or unmount is required. This allows easy addition of support for other filesystem types, such as MIT's Remote Virtual Disk (RVD) which has a programmatic interface via the commands @samp{rvdmount} and @samp{rvdunmount}. @noindent The following options must be specified: @table @code @cindex mount, mount option @cindex Mount option; mount @item mount the program which will perform the mount. @cindex unmount, mount option @cindex Mount option; unmount @item unmount the program which will perform the unmount. @end table The exit code from these two programs is interpreted as a Unix error code. As usual, exit code zero indicates success. To execute the program @i{Amd} splits the string on whitespace to create an array of substrings. Single quotes @samp{'} can be used to quote whitespace if that is required in an argument. There is no way to escape or change the quote character. To run the program @samp{rvdmount} with a host name and filesystem as arguments would be specified by @samp{mount:="/etc/rvdmount rvdmount fserver $@{path@}"}. The first element in the array is taken as the pathname of the program to execute. The other members of the array form the argument vector to be passed to the program, @dfn{including argument zero}. This means that the split string must have at least two elements. The program is directly executed by @i{Amd}, not via a shell. This means that scripts must begin with a @code{#!} interpreter specification. If a filesystem type is to be heavily used, it may be worthwhile adding a new filesystem type into @i{Amd}, but for most uses the program filesystem should suffice. When the program is run, standard input and standard error are inherited from the current values used by @i{Amd}. Standard output is a duplicate of standard error. The value specified with the ``-l'' command line option has no effect on standard error. @node Symbolic Link Filesystem, Symbolic Link Filesystem II, Program Filesystem, Filesystem Types @comment node-name, next, previous, up @section Symbolic Link Filesystem (@samp{type:=link}) @cindex Symbolic link filesystem @cindex Referencing part of the local name space @cindex Mounting part of the local name space @cindex How to reference part of the local name space @cindex link, filesystem type @cindex symlink, link filesystem type @cindex Filesystem type; link Each filesystem type creates a symbolic link to point from the volume name to the physical mount point. The @samp{link} filesystem does the same without any other side effects. This allows any part of the machines name space to be accessed via @i{Amd}. One common use for the symlink filesystem is @file{/homes} which can be made to contain an entry for each user which points to their (auto-mounted) home directory. Although this may seem rather expensive, it provides a great deal of administrative flexibility. @noindent The following option must be defined: @table @code @item fs The value of @var{fs} option specifies the destination of the link, as modified by the @var{sublink} option. If @var{sublink} is non-null, it is appended to @code{$@{fs@}}@code{/} and the resulting string is used as the target. @end table The @samp{link} filesystem can be though of as identical to the @samp{ufs} filesystem but without actually mounting anything. An example entry might be: @example jsp host==charm;type:=link;fs:=/home/charm;sublink:=jsp @end example which would return a symbolic link pointing to @file{/home/charm/jsp}. @node Symbolic Link Filesystem II, Automount Filesystem, Symbolic Link Filesystem, Filesystem Types @comment node-name, next, previous, up @section Symbolic Link Filesystem II (@samp{type:=linkx}) @cindex Symbolic link filesystem II @cindex Referencing an existing part of the local name space @cindex Mounting an existing part of the local name space @cindex How to reference an existing part of the local name space @cindex linkx, filesystem type @cindex symlink, linkx filesystem type @cindex Filesystem type; linkx The @samp{linkx} filesystem type is identical to @samp{link} with the exception that the target of the link must exist. Existence is checked with the @samp{lstat} system call. The @samp{linkx} filesystem type is particularly useful for wildcard map entries. In this case, a list of possible targets can be give and @i{Amd} will choose the first one which exists on the local machine. @node Automount Filesystem, Direct Automount Filesystem, Symbolic Link Filesystem II, Filesystem Types @comment node-name, next, previous, up @section Automount Filesystem (@samp{type:=auto}) @cindex Automount filesystem @cindex Map cache types @cindex Setting map cache parameters @cindex How to set map cache parameters @cindex How to start an indirect automount point @cindex auto, filesystem type @cindex Filesystem type; auto @cindex SIGHUP signal @cindex Map cache synchronising @cindex Synchronising the map cache @cindex Map cache options @cindex Regular expressions in maps The @dfn{auto} filesystem type creates a new automount point below an existing automount point. Top-level automount points appear as system mount points. An automount mount point can also appear as a sub-directory of an existing automount point. This allows some additional structure to be added, for example to mimic the mount tree of another machine. The following options may be specified: @table @code @cindex cache, mount option @cindex Mount option; cache @item cache specifies whether the data in this mount-map should be cached. The default value is @samp{none}, in which case no caching is done in order to conserve memory. However, better performance and reliability can be obtained by caching some or all of a mount-map. If the cache option specifies @samp{all}, the entire map is enumerated when the mount point is created. If the cache option specifies @samp{inc}, caching is done incrementally as and when data is required. Some map types do not support cache mode @samp{all}, in which case @samp{inc} is used whenever @samp{all} is requested. Caching can be entirely disabled by using cache mode @samp{none}. If the cache option specifies @samp{regexp} then the entire map will be enumerated and each key will be treated as an egrep-style regular expression. The order in which a cached map is searched does not correspond to the ordering in the source map so the regular expressions should be mutually exclusive to avoid confusion. Each mount map type has a default cache type, usually @samp{inc}, which can be selected by specifying @samp{mapdefault}. The cache mode for a mount map can only be selected on the command line. Starting @i{Amd} with the command: @example amd /homes hesiod.homes -cache:=inc @end example will cause @samp{/homes} to be automounted using the @dfn{Hesiod} name server with local incremental caching of all successfully resolved names. All cached data is forgotten whenever @i{Amd} receives a @samp{SIGHUP} signal and, if cache @samp{all} mode was selected, the cache will be reloaded. This can be used to inform @i{Amd} that a map has been updated. In addition, whenever a cache lookup fails and @i{Amd} needs to examine a map, the map's modify time is examined. If the cache is out of date with respect to the map then it is flushed as if a @samp{SIGHUP} had been received. An additional option (@samp{sync}) may be specified to force @i{Amd} to check the map's modify time whenever a cached entry is being used. For example, an incremental, synchronised cache would be created by the following command: @example amd /homes hesiod.homes -cache:=inc,sync @end example @item fs specifies the name of the mount map to use for the new mount point. Arguably this should have been specified with the @code{$@{rfs@}} option but we are now stuck with it due to historical accident. @c %If the string @samp{.} is used then the same map is used; @c %in addition the lookup prefix is set to the name of the mount point followed @c %by a slash @samp{/}. @c %This is the same as specifying @samp{fs:=\$@{map@};pref:=\$@{key@}/}. @c @item pref alters the name that is looked up in the mount map. If @code{$@{pref@}}, the @dfn{prefix}, is non-null then it is prepended to the name requested by the kernel @dfn{before} the map is searched. @end table The server @samp{dylan.doc.ic.ac.uk} has two user disks: @samp{/dev/dsk/2s0} and @samp{/dev/dsk/5s0}. These are accessed as @samp{/home/dylan/dk2} and @samp{/home/dylan/dk5} respectively. Since @samp{/home} is already an automount point, this naming is achieved with the following map entries:@refill @example dylan type:=auto;fs:=$@{map@};pref:=$@{key@}/ dylan/dk2 type:=ufs;dev:=/dev/dsk/2s0 dylan/dk5 type:=ufs;dev:=/dev/dsk/5s0 @end example @node Direct Automount Filesystem, Union Filesystem, Automount Filesystem, Filesystem Types @comment node-name, next, previous, up @section Direct Automount Filesystem (@samp{type:=direct}) @cindex Direct automount filesystem @cindex How to start a direct automount point @cindex direct, filesystem type @cindex Filesystem type; direct The @dfn{direct} filesystem is almost identical to the automount filesystem. Instead of appearing to be a directory of mount points, it appears as a symbolic link to a mounted filesystem. The mount is done at the time the link is accessed. @xref{Automount Filesystem}, for a list of required options. Direct automount points are created by specifying the @samp{direct} filesystem type on the command line: @example amd ... /usr/man auto.direct -type:=direct @end example where @samp{auto.direct} would contain an entry such as: @example usr/man -type:=nfs;rfs:=/usr/man \ rhost:=man-server1 rhost:=man-server2 @end example In this example, @samp{man-server1} and @samp{man-server2} are file servers which export copies of the manual pages. Note that the key which is looked up is the name of the automount point without the leading @samp{/}. @node Union Filesystem, Error Filesystem, Direct Automount Filesystem, Filesystem Types @comment node-name, next, previous, up @section Union Filesystem (@samp{type:=union}) @cindex Union filesystem @cindex union, filesystem type @cindex Filesystem type; union The @dfn{union} filesystem type allows the contents of several directories to be merged and made visible in a single directory. This can be used to overcome one of the major limitations of the Unix mount mechanism which only allows complete directories to be mounted. For example, supposing @file{/tmp} and @file{/var/tmp} were to be merged into a new directory called @file{/mtmp}, with files in @file{/var/tmp} taking precedence. The following command could be used to achieve this effect: @example amd ... /mtmp union:/tmp:/var/tmp -type:=union @end example Currently, the unioned directories must @emph{not} be automounted. That would cause a deadlock. This seriously limits the current usefulness of this filesystem type and the problem will be addressed in a future release of @i{Amd}. Files created in the union directory are actually created in the last named directory. This is done by creating a wildcard entry which points to the correct directory. The wildcard entry is visible if the union directory is listed, so allowing you to see which directory has priority. The files visible in the union directory are computed at the time @i{Amd} is started, and are not kept uptodate with respect to the underlying directories. Similarly, if a link is removed, for example with the @samp{rm} command, it will be lost forever. @node Error Filesystem, Top-level Filesystem, Union Filesystem, Filesystem Types @comment node-name, next, previous, up @section Error Filesystem (@samp{type:=error}) @cindex Error filesystem @cindex error, filesystem type @cindex Filesystem type; error The @dfn{error} filesystem type is used internally as a catch-all in the case where none of the other filesystems was selected, or some other error occurred. Lookups and mounts always fail with ``No such file or directory''. All other operations trivially succeed. The error filesystem is not directly accessible. @node Top-level Filesystem, Root Filesystem, Error Filesystem, Filesystem Types @comment node-name, next, previous, up @section Top-level Filesystem (@samp{type:=toplvl}) @cindex Top level filesystem @cindex toplvl, filesystem type @cindex Filesystem type; toplvl The @dfn{toplvl} filesystems is derived from the @samp{auto} filesystem and is used to mount the top-level automount nodes. Requests of this type are automatically generated from the command line arguments and can also be passed in by using the ``-M'' option of the @dfn{Amq} command. @node Root Filesystem, Inheritance Filesystem, Top-level Filesystem, Filesystem Types @comment node-name, next, previous, up @section Root Filesystem @cindex Root filesystem @cindex root, filesystem type @cindex Filesystem type; root The @dfn{root} (@samp{type:=root}) filesystem type acts as an internal placeholder onto which @i{Amd} can pin @samp{toplvl} mounts. Only one node of this type need ever exist and one is created automatically during startup. The effect of creating a second root node is undefined. @node Inheritance Filesystem, , Root Filesystem, Filesystem Types @comment node-name, next, previous, up @section Inheritance Filesystem @cindex Inheritance filesystem @cindex Nodes generated on a restart @cindex inherit, filesystem type @cindex Filesystem type; inherit The @dfn{inheritance} (@samp{type:=inherit}) filesystem is not directly accessible. Instead, internal mount nodes of this type are automatically generated when @i{Amd} is started with the ``-r'' option. At this time the system mount table is scanned to locate any filesystems which are already mounted. If any reference to these filesystems is made through @i{Amd} then instead of attempting to mount it, @i{Amd} simulates the mount and @dfn{inherits} the filesystem. This allows a new version of @i{Amd} to be installed on a live system simply by killing the old daemon with @code{SIGTERM} and starting the new one.@refill This filesystem type is not generally visible externally, but it is possible that the output from @samp{amq -m} may list @samp{inherit} as the filesystem type. This happens when an inherit operation cannot be completed for some reason, usually because a fileserver is down. @node Run-time Administration, Examples, Filesystem Types, Top @comment node-name, next, previous, up @chapter Run-time Administration @cindex Run-time administration @cindex Amq command @menu * Starting Amd:: * Stopping Amd:: * Controlling Amd:: @end menu @node Starting Amd, Stopping Amd, Run-time Administration, Run-time Administration @comment node-name, next, previous, up @section Starting @i{Amd} @cindex Starting Amd @cindex Additions to /etc/rc.local @cindex /etc/rc.local additions @cindex /etc/amd.start @i{Amd} is best started from @samp{/etc/rc.local}: @example if [ -f /etc/amd.start ]; then sh /etc/amd.start; (echo -n ' amd') >/dev/console fi @end example @noindent The shell script, @samp{amd.start}, contains: @example #!/bin/sh - PATH=/etc:/bin:/usr/bin:/usr/ucb:$PATH export PATH # # Either name of logfile or "syslog" # LOGFILE=syslog #LOGFILE=/var/log/amd # # Figure out whether domain name is in host name # If the hostname is just the machine name then # pass in the name of the local domain so that the # hostnames in the map are domain stripped correctly. # case `hostname` in *.*) dmn= ;; *) dmn='-d doc.ic.ac.uk' esac # # Zap earlier log file # case "$LOGFILE" in */*) mv "$LOGFILE" "$LOGFILE"- > "$LOGFILE" ;; syslog) : nothing ;; esac cd /usr/sbin # # -r restart # -d dmn local domain # -w wait wait between unmount attempts # -l log logfile or "syslog" # eval ./amd -r $dmn -w 240 -l "$LOGFILE" \ /homes amd.homes -cache:=inc \ /home amd.home -cache:=inc \ /vol amd.vol -cache:=inc \ /n amd.net -cache:=inc @end example If the list of automount points and maps is contained in a file or NIS map it is easily incorporated onto the command line: @example ... eval ./amd -r $dmn -w 240 -l "$LOGFILE" `ypcat -k auto.master` @end example @node Stopping Amd, Controlling Amd, Starting Amd, Run-time Administration @comment node-name, next, previous, up @section Stopping @i{Amd} @cindex Stopping Amd @cindex SIGTERM signal @cindex SIGINT signal @i{Amd} stops in response to two signals. @table @samp @item SIGTERM causes the top-level automount points to be unmounted and then @i{Amd} to exit. Any automounted filesystems are left mounted. They can be recovered by restarting @i{Amd} with the ``-r'' command line option.@refill @item SIGINT causes @i{Amd} to attempt to unmount any filesystems which it has automounted, in addition to the actions of @samp{SIGTERM}. This signal is primarly used for debugging.@refill @end table Actions taken for other signals are undefined. @node Controlling Amd, , Stopping Amd, Run-time Administration @comment node-name, next, previous, up @section Controlling @i{Amd} @cindex Controlling Amd @cindex Discovering what is going on at run-time @cindex Listing currently mounted filesystems It is sometimes desirable or necessary to exercise external control over some of @i{Amd}'s internal state. To support this requirement, @i{Amd} implements an RPC interface which is used by the @dfn{Amq} program. A variety of information is available. @i{Amq} generally applies an operation, specified by a single letter option, to a list of mount points. The default operation is to obtain statistics about each mount point. This is similar to the output shown above but includes information about the number and type of accesses to each mount point. @menu * Amq default:: Default command behaviour. * Amq -f option:: Flushing the map cache. * Amq -h option:: Controlling a non-local host. * Amq -m option:: Obtaining mount statistics. * Amq -M-option:: Mounting a volume. * Amq -s option:: Obtaining global statistics. * Amq -u option:: Forcing volumes to time out. * Amq -v option:: Version information. * Other Amq options:: Three other special options. @end menu @node Amq default, Amq -f option, Controlling Amd, Controlling Amd @comment node-name, next, previous, up @subsection @i{Amq} default information With no arguments, @dfn{Amq} obtains a brief list of all existing mounts created by @i{Amd}. This is different from the list displayed by @b{df}(1) since the latter only includes system mount points. @noindent The output from this option includes the following information: @itemize @bullet @item the automount point, @item the filesystem type, @item the mount map or mount information, @item the internal, or system mount point. @end itemize @noindent For example: @example / root "root" sky:(pid75) /homes toplvl /usr/local/etc/amd.homes /homes /home toplvl /usr/local/etc/amd.home /home /homes/jsp nfs charm:/home/charm /a/charm/home/charm/jsp /homes/phjk nfs toytown:/home/toytown /a/toytown/home/toytown/ai/phjk @end example @noindent If an argument is given then statistics for that volume name will be output. For example: @example What Uid Getattr Lookup RdDir RdLnk Statfs Mounted@@ /homes 0 1196 512 22 0 30 90/09/14 12:32:55 /homes/jsp 0 0 0 0 1180 0 90/10/13 12:56:58 @end example @table @code @item What the volume name. @item Uid ignored. @item Getattr the count of NFS @dfn{getattr} requests on this node. This should only be non-zero for directory nodes. @item Lookup the count of NFS @dfn{lookup} requests on this node. This should only be non-zero for directory nodes. @item RdDir the count of NFS @dfn{readdir} requests on this node. This should only be non-zero for directory nodes. @item RdLnk the count of NFS @dfn{readlink} requests on this node. This should be zero for directory nodes. @item Statfs the could of NFS @dfn{statfs} requests on this node. This should only be non-zero for top-level automount points. @item Mounted@@ the date and time the volume name was first referenced. @end table @node Amq -f option, Amq -h option, Amq default, Controlling Amd @comment node-name, next, previous, up @subsection @i{Amq} -f option @cindex Flushing the map cache @cindex Map cache, flushing The ``-f'' option causes @i{Amd} to flush the internal mount map cache. This is useful for Hesiod maps since @i{Amd} will not automatically notice when they have been updated. The map cache can also be synchronised with the map source by using the @samp{sync} option (@pxref{Automount Filesystem}).@refill @node Amq -h option, Amq -m option, Amq -f option, Controlling Amd @comment node-name, next, previous, up @subsection @i{Amq} -h option @cindex Querying an alternate host By default the local host is used. In an HP-UX cluster the root server is used since that is the only place in the cluster where @i{Amd} will be running. To query @i{Amd} on another host the ``-h'' option should be used. @node Amq -m option, Amq -M-option, Amq -h option, Controlling Amd @comment node-name, next, previous, up @subsection @i{Amq} -m option The ``-m'' option displays similar information about mounted filesystems, rather than automount points. The output includes the following information: @itemize @bullet @item the mount information, @item the mount point, @item the filesystem type, @item the number of references to this filesystem, @item the server hostname, @item the state of the file server, @item any error which has occured. @end itemize For example: @example "root" truth:(pid602) root 1 localhost is up hesiod.home /home toplvl 1 localhost is up hesiod.vol /vol toplvl 1 localhost is up hesiod.homes /homes toplvl 1 localhost is up amy:/home/amy /a/amy/home/amy nfs 5 amy is up swan:/home/swan /a/swan/home/swan nfs 0 swan is up (Permission denied) ex:/home/ex /a/ex/home/ex nfs 0 ex is down @end example When the reference count is zero the filesystem is not mounted but the mount point and server information is still being maintained by @i{Amd}. @node Amq -M-option, Amq -s option, Amq -m option, Controlling Amd @comment node-name, next, previous, up @subsection @i{Amq} -M option The ``-M'' option passes a new map entry to @i{Amd} and waits for it to be evaluated, possibly causing a mount. For example, the following command would cause @samp{/home/toytown} on host @samp{toytown} to be mounted locally on @samp{/mnt/toytown}. @example amq -M '/mnt/toytown type:=nfs;rfs:=/home/toytown;rhost:=toytown;fs:=$@{key@}' @end example @i{Amd} applies some simple security checks before allowing this operation. The check tests whether the incoming request is from a privileged UDP port on the local machine. ``Permission denied'' is returned if the check fails. A future release of @i{Amd} will include code to allow the @b{mount}(8) command to mount automount points: @example mount -t amd /vol hesiod.vol @end example This will then allow @i{Amd} to be controlled from the standard system filesystem mount list. @node Amq -s option, Amq -u option, Amq -M-option, Controlling Amd @comment node-name, next, previous, up @subsection @i{Amq} -s option @cindex Global statistics @cindex Statistics The ``-s'' option displays global statistics. If any other options are specified or any filesystems named then this option is ignored. For example: @example requests stale mount mount unmount deferred fhandles ok failed failed 1054 1 487 290 7017 @end example @table @samp @item Deferred requests are those for which an immediate reply could not be constructed. For example, this would happen if a background mount was required. @item Stale filehandles counts the number of times the kernel passes a stale filehandle to @i{Amd}. Large numbers indicate problems. @item Mount ok counts the number of automounts which were successful. @item Mount failed counts the number of automounts which failed. @item Unmount failed counts the number of times a filesystem could not be unmounted. Very large numbers here indicate that the time between unmount attempts should be increased. @end table @node Amq -u option, Amq -v option, Amq -s option, Controlling Amd @comment node-name, next, previous, up @subsection @i{Amq} -u option @cindex Forcing filesystem to time out @cindex Unmounting a filesystem The ``-u'' option causes the time-to-live interval of the named mount points to be expired, thus causing an unmount attempt. This is the only safe way to unmount an automounted filesystem. It is not possible to unmount a filesystem which has been mounted with the @samp{nounmount} flag. @c The ``-H'' option informs @i{Amd} that the specified mount point has hung - @c as if its keepalive timer had expired. @node Amq -v option, Other Amq options, Amq -u option, Controlling Amd @comment node-name, next, previous, up @subsection @i{Amq} -v option @cindex Version information at run-time The ``-v'' option displays the version of @i{Amd} in a similar way to @i{Amd}'s ``-v'' option. @node Other Amq options, , Amq -v option, Controlling Amd @comment node-name, next, previous, up @subsection Other @i{Amq} options Three other operations are implemented. These modify the state of @i{Amd} as a whole, rather than any particular filesystem. The ``-l'', ``-x'' and ``-D'' options have exactly the same effect as @i{Amd}'s corresponding command line options. The ``-l'' option is rejected by @i{Amd} in the current version for obvious security reasons. When @i{Amd} receives a ``-x''flag it limits the log options being modified to those which were not enabled at startup. This prevents a user turning @emph{off} any logging option which was specified at startup, though any which have been turned off since then can still be turned off. The ``-D'' option has a similar behaviour. @node Examples, Internals, Run-time Administration, Top @comment node-name, next, previous, up @chapter Examples @menu * User Filesystems:: * Home Directories:: * Architecture Sharing:: * Wildcard Names:: * rwho servers:: * /vol:: @end menu @node User Filesystems, Home Directories, Examples, Examples @comment node-name, next, previous, up @section User Filesystems @cindex User filesystems @cindex Mounting user filesystems With more than one fileserver, the directories most frequently cross-mounted are those containing user home directories. A common convention used at Imperial College is to mount the user disks under @t{/home/}@i{machine}. Typically, the @samp{/etc/fstab} file contained a long list of entries such as: @example @i{machine}:/home/@i{machine} /home/@i{machine} nfs ... @end example for each fileserver on the network. There are numerous problems with this system. The mount list can become quite large and some of the machines may be down when a system is booted. When a new fileserver is installed, @samp{/etc/fstab} must be updated on every machine, the mount directory created and the filesystem mounted. In many environments most people use the same few workstations, but it is convenient to go to a colleague's machine and access your own files. When a server goes down, it can cause a process on a client machine to hang. By minimising the mounted filesystems to only include those actively being used, there is less chance that a filesystem will be mounted when a server goes down. The following is a short extract from a map taken from a research fileserver at Imperial College. Note the entry for @samp{localhost} which is used for users such as the operator (@samp{opr}) who have a home directory on most machine as @samp{/home/localhost/opr}. @example /defaults opts:=rw,intr,grpid,nosuid charm host!=$@{key@};type:=nfs;rhost:=$@{key@};rfs:=/home/$@{key@} \ host==$@{key@};type:=ufs;dev:=/dev/xd0g # ... # localhost type:=link;fs:=$@{host@} ... # # dylan has two user disks so have a # top directory in which to mount them. # dylan type:=auto;fs:=$@{map@};pref:=$@{key@}/ # dylan/dk2 host!=dylan;type:=nfs;rhost:=dylan;rfs:=/home/$@{key@} \ host==dylan;type:=ufs;dev:=/dev/dsk/2s0 # dylan/dk5 host!=dylan;type:=nfs;rhost:=dylan;rfs:=/home/$@{key@} \ host==dylan;type:=ufs;dev:=/dev/dsk/5s0 ... # toytown host!=$@{key@};type:=nfs;rhost:=$@{key@};rfs:=/home/$@{key@} \ host==$@{key@};type:=ufs;dev:=/dev/xy1g ... # zebedee host!=$@{key@};type:=nfs;rhost:=$@{key@};rfs:=/home/$@{key@} \ host==$@{key@};type:=ufs;dev:=/dev/dsk/1s0 # # Just for access... # gould type:=auto;fs:=$@{map@};pref:=$@{key@}/ gould/staff host!=gould;type:=nfs;rhost:=gould;rfs:=/home/$@{key@} # gummo host!=$@{key@};type:=nfs;rhost:=$@{key@};rfs:=/home/$@{key@} ... @end example This map is shared by most of the machines listed so on those systems any of the user disks is accessible via a consistent name. @i{Amd} is started with the following command @example amd /home amd.home @end example Note that when mounting a remote filesystem, the @dfn{automounted} mount point is referenced, so that the filesystem will be mounted if it is not yet (at the time the remote @samp{mountd} obtains the file handle). @node Home Directories, Architecture Sharing, User Filesystems, Examples @comment node-name, next, previous, up @section Home Directories @cindex Home directories @cindex Example of mounting home directories @cindex Mount home directories One convention for home directories is to locate them in @samp{/homes} so user @samp{jsp}'s home directory is @samp{/homes/jsp}. With more than a single fileserver it is convenient to spread user files across several machines. All that is required is a mount-map which converts login names to an automounted directory. Such a map might be started by the command: @example amd /homes amd.homes @end example where the map @samp{amd.homes} contained the entries: @example /defaults type:=link # All the entries are of type:=link jsp fs:=/home/charm/jsp njw fs:=/home/dylan/dk5/njw ... phjk fs:=/home/toytown/ai/phjk sjv fs:=/home/ganymede/sjv @end example Whenever a login name is accessed in @samp{/homes} a symbolic link appears pointing to the real location of that user's home directory. In this example, @samp{/homes/jsp} would appear to be a symbolic link pointing to @samp{/home/charm/jsp}. Of course, @samp{/home} would also be an automount point. This system causes an extra level of symbolic links to be used. Although that turns out to be relatively inexpensive, an alternative is to directly mount the required filesystems in the @samp{/homes} map. The required map is simple, but long, and its creation is best automated. The entry for @samp{jsp} could be: @example jsp -sublink:=$@{key@};rfs:=/home/charm \ host==charm;type:=ufs;dev:=/dev/xd0g \ host!=charm;type:=nfs;rhost:=charm @end example This map can become quite big if it contains a large number of entries. By combining two other features of @i{Amd} it can be greatly simplified. First the UFS partitions should be mounted under the control of @samp{/etc/fstab}, taking care that they are mounted in the same place that @i{Amd} would have automounted them. In most cases this would be something like @samp{/a/@dfn{host}/home/@dfn{host}} and @samp{/etc/fstab} on host @samp{charm} would have a line:@refill @example /dev/xy0g /a/charm/home/charm 4.2 rw,nosuid,grpid 1 5 @end example The map can then be changed to: @example /defaults type:=nfs;sublink:=$@{key@};opts:=rw,intr,nosuid,grpid jsp rhost:=charm;rfs:=/home/charm njw rhost:=dylan;rfs:=/home/dylan/dk5 ... phjk rhost:=toytown;rfs:=/home/toytown;sublink:=ai/$@{key@} sjv rhost:=ganymede;rfs:=/home/ganymede @end example This map operates as usual on a remote machine (@i{ie} @code{$@{host@}} not equal to @code{$@{rhost@}}). On the machine where the filesystem is stored (@i{ie} @code{$@{host@}} equal to @code{$@{rhost@}}), @i{Amd} will construct a local filesystem mount point which corresponds to the name of the locally mounted UFS partition. If @i{Amd} is started with the ``-r'' option then instead of attempting an NFS mount, @i{Amd} will simply inherit the UFS mount (@pxref{Inheritance Filesystem}). If ``-r'' is not used then a loopback NFS mount will be made. This type of mount is known to cause a deadlock on many systems. @node Architecture Sharing, Wildcard Names, Home Directories, Examples @comment node-name, next, previous, up @section Architecture Sharing @cindex Architecture sharing @cindex Sharing a fileserver between architectures @cindex Architecture dependent volumes @c %At the moment some of the research machines have sets of software @c %mounted in @samp{/vol}. This contains subdirectories for \TeX, @c %system sources, local sources, prolog libraries and so on. Often a filesystem will be shared by machines of different architectures. Separate trees can be maintained for the executable images for each architecture, but it may be more convenient to have a shared tree, with distinct subdirectories. A shared tree might have the following structure on the fileserver (called @samp{fserver} in the example): @example local/tex local/tex/fonts local/tex/lib local/tex/bin local/tex/bin/sun3 local/tex/bin/sun4 local/tex/bin/hp9000 ... @end example In this example, the subdirectories of @samp{local/tex/bin} should be hidden when accessed via the automount point (conventionally @samp{/vol}). A mount-map for @samp{/vol} to achieve this would look like: @example /defaults sublink:=$@{/key@};rhost:=fserver;type:=link tex type:=auto;fs:=$@{map@};pref:=$@{key@}/ tex/fonts host!=fserver;type:=nfs;rfs:=/vol/tex \ host==fserver;fs:=/usr/local/tex tex/lib host!=fserver;type:=nfs;rfs:=/vol/tex \ host==fserver;fs:=/usr/local/tex tex/bin -sublink:=$@{/key@}/$@{arch@} host!=fserver;type:=nfs;rfs:=/vol/tex \ host:=fserver;fs:=/usr/local/tex @end example When @samp{/vol/tex/bin} is referenced, the current machine architecture is automatically appended to the path by the @code{$@{sublink@}} variable. This means that users can have @samp{/vol/tex/bin} in their @samp{PATH} without concern for architecture dependencies. @node Wildcard Names, rwho servers, Architecture Sharing, Examples @comment node-name, next, previous, up @section Wildcard names & Replicated Servers By using the wildcard facility, @i{Amd} can @dfn{overlay} an existing directory with additional entries. The system files are usually mounted under @samp{/usr}. If instead @i{Amd} is mounted on @samp{/usr}, additional names can be overlayed to augment or replace names in the ``master'' @samp{/usr}. A map to do this would have the form: @example local type:=auto;fs:=local-map share type:=auto;fs:=share-map * -type:=nfs;rfs:=/export/exec/$@{arch@};sublink:="$@{key@}" \ rhost:=fserv1 rhost:=fserv2 rhost:=fserv3 @end example Note that the assignment to @code{$@{sublink@}} is surrounded by double quotes to prevent the incoming key from causing the map to be misinterpreted. This map has the effect of directing any access to @samp{/usr/local} or @samp{/usr/share} to another automount point. In this example, it is assumed that the @samp{/usr} files are replicated on three fileservers: @samp{fserv1}, @samp{fserv2} and @samp{fserv3}. For any references other than to @samp{local} and @samp{share} one of the servers is used and a symbolic link to @t{$@{autodir@}/$@{rhost@}/export/exec/$@{arch@}/@i{whatever}} is returned once an appropriate filesystem has been mounted.@refill @node rwho servers, /vol, Wildcard Names, Examples @comment node-name, next, previous, up @section @samp{rwho} servers @cindex rwho servers @cindex Architecture specific mounts @cindex Example of architecture specific mounts The @samp{/usr/spool/rwho} directory is a good candidate for automounting. For efficiency reasons it is best to capture the rwho data on a small number of machines and then mount that information onto a large number of clients. The data written into the rwho files is byte order dependent so only servers with the correct byte ordering can be used by a client: @example /defaults type:=nfs usr/spool/rwho -byte==little;rfs:=/usr/spool/rwho \ rhost:=vaxA rhost:=vaxB \ || -rfs:=/usr/spool/rwho \ rhost:=sun4 rhost:=hp300 @end example @node /vol, , rwho servers, Examples @comment node-name, next, previous, up @section @samp{/vol} @cindex /vol @cindex Catch-all mount point @cindex Generic volume name @samp{/vol} is used as a catch-all for volumes which do not have other conventional names. Below is part of the @samp{/vol} map for the domain @samp{doc.ic.ac.uk}. The @samp{r+d} tree is used for new or experimental software that needs to be available everywhere without installing it on all the fileservers. Users wishing to try out the new software then simply include @samp{/vol/r+d/@{bin,ucb@}} in their path.@refill The main tree resides on one host @samp{gould.doc.ic.ac.uk}, which has different @samp{bin}, @samp{etc}, @samp{lib} and @samp{ucb} sub-directories for each machine architecture. For example, @samp{/vol/r+d/bin} for a Sun-4 would be stored in the sub-directory @samp{bin/sun4} of the filesystem @samp{/usr/r+d}. When it was accessed a symbolic link pointing to @samp{/a/gould/usr/r+d/bin/sun4} would be returned.@refill @example /defaults type:=nfs;opts:=rw,grpid,nosuid,intr,soft wp -opts:=rw,grpid,nosuid;rhost:=charm \ host==charm;type:=link;fs:=/usr/local/wp \ host!=charm;type:=nfs;rfs:=/vol/wp ... # src -opts:=rw,grpid,nosuid;rhost:=charm \ host==charm;type:=link;fs:=/usr/src \ host!=charm;type:=nfs;rfs:=/vol/src # r+d type:=auto;fs:=$@{map@};pref:=r+d/ # per architecture bin,etc,lib&ucb... r+d/bin rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}/$@{arch@} r+d/etc rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}/$@{arch@} r+d/include rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@} r+d/lib rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}/$@{arch@} r+d/man rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@} r+d/src rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@} r+d/ucb rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}/$@{arch@} # hades pictures pictures -opts:=rw,grpid,nosuid;rhost:=thpfs \ host==thpfs;type:=link;fs:=/nbsd/pictures \ host!=thpfs;type:=nfs;rfs:=/nbsd;sublink:=pictures # hades tools hades -opts:=rw,grpid,nosuid;rhost:=thpfs \ host==thpfs;type:=link;fs:=/nbsd/hades \ host!=thpfs;type:=nfs;rfs:=/nbsd;sublink:=hades # bsd tools for hp. bsd -opts:=rw,grpid,nosuid;arch==hp9000;rhost:=thpfs \ host==thpfs;type:=link;fs:=/nbsd/bsd \ host!=thpfs;type:=nfs;rfs:=/nbsd;sublink:=bsd @end example @node Internals, Acknowledgements & Trademarks, Examples, Top @comment node-name, next, previous, up @chapter Internals @menu * Log Messages:: @end menu @node Log Messages, , Internals, Internals @comment node-name, next, previous, up @section Log Messages In the following sections a brief explanation is given of some of the log messages made by @i{Amd}. Where the message is in @samp{typewriter} font, it corresponds exactly to the message produced by @i{Amd}. Words in @dfn{italic} are replaced by an appropriate string. Variables, @code{$@{var@}}, indicate that the value of the appropriate variable is output. Log messages are either sent direct to a file, or logged via the @b{syslog}(3) mechanism. Messages are logged with facility @samp{LOG_DAEMON} when using @b{syslog}(3). In either case, entries in the file are of the form: @example @i{date-string} @i{hostname} @t{amd[}@i{pid}@t{]} @i{message} @end example @menu * Fatal errors:: * Info messages:: @end menu @node Fatal errors, Info messages, Log Messages, Log Messages @comment node-name, next, previous, up @subsection Fatal errors @i{Amd} attempts to deal with unusual events. Whenever it is not possible to deal with such an error, @i{Amd} will log an appropriate message and, if it cannot possibly continue, will either exit or abort. These messages are selected by @samp{-x fatal} on the command line. When @b{syslog}(3) is being used, they are logged with level @samp{LOG_FATAL}. Even if @i{Amd} continues to operate it is likely to remain in a precarious state and should be restarted at the earliest opportunity. @table @asis @item @t{Attempting to inherit not-a-filesystem} The prototype mount point created during a filesystem restart did not contain a reference to the restarted filesystem. This erorr ``should never happen''. @item @t{Can't bind to domain "@i{NIS-domain}"} A specific NIS domain was requested on the command line, but no server for that domain is available on the local net. @item @t{Can't determine IP address of this host (@i{hostname})} When @i{Amd} starts it determines its own IP address. If this lookup fails then @i{Amd} cannot continue. The hostname it looks up is that obtained returned by @b{gethostname}(2) system call. @item @t{Can't find root file handle for @i{automount point}} @i{Amd} creates its own file handles for the automount points. When it mounts itself as a server, it must pass these file handles to the local kernel. If the filehandle is not obtainable the mount point is ignored. This error ``should never happen''. @item @t{Must be root to mount filesystems (euid = @i{euid})} To prevent embarrassment, @i{Amd} makes sure it has appropriate system privileges. This amounts to having an euid of 0. The check is made after argument processing complete to give non-root users a chance to access the ``-v'' option. @item @t{No work to do - quitting} No automount points were given on the command line and so there is no work to do. @item @t{Out of memory in realloc} While attempting to realloc some memory, the memory space available to @i{Amd} was exhausted. This is an unrecoverable error. @item @t{Out of memory} While attempting to malloc some memory, the memory space available to @i{Amd} was exhausted. This is an unrecoverable error. @item @t{cannot create rpc/udp service} Either the NFS or AMQ endpoint could not be created. @item @t{gethostname:} @i{description} The @b{gethostname}(2) system call failed during startup. @item @t{host name is not set} The @b{gethostname}(2) system call returned a zero length host name. This can happen if @i{Amd} is started in single user mode just after booting the system. @item @t{ifs_match called!} An internal error occurred while restarting a pre-mounted filesystem. This error ``should never happen''. @item @t{mount_afs:} @i{description} An error occured while @i{Amd} was mounting itself. @item @t{run_rpc failed} Somehow the main NFS server loop failed. This error ``should never happen''. @item @t{unable to free rpc arguments in amqprog_1} The incoming arguments to the AMQ server could not be free'ed. @item @t{unable to free rpc arguments in nfs_program_1} The incoming arguments to the NFS server could not be free'ed. @item @t{unable to register (AMQ_PROGRAM, AMQ_VERSION, udp)} The AMQ server could not be registered with the local portmapper or the internal RPC dispatcher. @item @t{unable to register (NFS_PROGRAM, NFS_VERSION, 0)} The NFS server could not be registered with the internal RPC dispatcher. @end table @node Info messages, , Fatal errors, Log Messages @comment node-name, next, previous, up @subsection Info messages @i{Amd} generates information messages to record state changes. These messages are selected by @samp{-x info} on the command line. When @b{syslog}(3) is being used, they are logged with level @samp{LOG_INFO}. The messages listed below can be generated and are in a format suitable for simple statistical analysis. @dfn{mount-info} is the string that is displayed by @dfn{Amq} in its mount information column and placed in the system mount table. @table @asis @item @t{mount of "@t{$@{@i{path}@}}" on @t{$@{@i{fs}@}} timed out} Attempts to mount a filesystem for the given automount point have failed to complete within 30 seconds. @item @t{"@t{$@{@i{path}@}}" forcibly timed out} An automount point has been timed out by the @i{Amq} command. @item @t{restarting @i{mount-info} on @t{$@{@i{fs}@}}} A pre-mounted file system has been noted. @item @t{"@t{$@{@i{path}@}}" has timed out} No access to the automount point has been made within the timeout period. @item @t{file server @t{$@{@i{rhost}@}} is down - timeout of "@t{$@{@i{path}@}}" ignored} An automount point has timed out, but the corresponding file server is known to be down. This message is only produced once for each mount point for which the server is down. @item @t{Re-synchronizing cache for map @t{$@{@i{map}@}}} The named map has been modified and the internal cache is being re-synchronized. @item @t{Filehandle denied for "@t{$@{@i{rhost}@}}:@t{$@{@i{rfs}@}}"} The mount daemon refused to return a file handle for the requested filesystem. @item @t{Filehandle error for "$@{@i{rhost}@}:$@{@i{rfs}@}":} @i{description} The mount daemon gave some other error for the requested filesystem. @item @t{file server @t{$@{@i{rhost}@}} type nfs starts up} A new NFS file server has been referenced and is known to be up. @item @t{file server @t{$@{@i{rhost}@}} type nfs starts down} A new NFS file server has been referenced and is known to be down. @item @t{file server @t{$@{@i{rhost}@}} type nfs is up} An NFS file server that was previously down is now up. @item @t{file server @t{$@{@i{rhost}@}} type nfs is down} An NFS file server that was previously up is now down. @item @t{Finishing with status @i{exit-status}} @i{Amd} is about to exit with the given exit status. @item @t{@i{mount-info} mounted fstype @t{$@{@i{type}@}} on @t{$@{@i{fs}@}}} A new file system has been mounted. @item @t{@i{mount-info} restarted fstype @t{$@{@i{type}@}} on @t{$@{@i{fs}@}}} @i{Amd} is using a pre-mounted filesystem to satisfy a mount request. @item @t{@i{mount-info} unmounted fstype @t{$@{@i{type}@}} from @t{$@{@i{fs}@}}} A file system has been unmounted. @item @t{@i{mount-info} unmounted fstype @t{$@{@i{type}@}} from @t{$@{@i{fs}@}} link @t{$@{@i{fs}@}}/@t{$@{@i{sublink}@}}} A file system of which only a sub-directory was in use has been unmounted. @end table @node Acknowledgements & Trademarks, Index, Internals, Top @comment node-name, next, previous, up @unnumbered Acknowledgements & Trademarks Thanks to the Formal Methods Group at Imperial College for suffering patiently while @i{Amd} was being developed on their machines. Thanks to the many people who have helped with the development of @i{Amd}, especially Piete Brooks at the Cambridge University Computing Lab for many hours of testing, experimentation and discussion. @itemize @bullet @item @b{DEC}, @b{VAX} and @b{Ultrix} are registered trademarks of Digital Equipment Corporation. @item @b{AIX} and @b{IBM} are registered trademarks of International Business Machines Corporation. @item @b{Sun}, @b{NFS} and @b{SunOS} are registered trademarks of Sun Microsystems, Inc. @item @b{Unix} is a registered trademark of AT&T Unix Systems Laboratories in the USA and other countries. @end itemize @node Index, Intro, Acknowledgements & Trademarks, Top @comment node-name, next, previous, up @unnumbered Index @printindex cp @contents @bye