3 afsd - Initializes the Cache Manager and starts related daemons
10 B<afsd> [B<-afsdb>] [B<-backuptree>]
11 S<<< [B<-biods> <I<number of bkg I/O daemons (aix vm)>>] >>>
12 S<<< [B<-blocks> <I<1024 byte blocks in cache>>] >>>
13 S<<< [B<-cachedir> <I<cache directory>>] >>>
14 S<<< [B<-chunksize> <I<log(2) of chunk size>>] >>>
15 S<<< [B<-confdir> <I<configuration directory>>] >>>
16 S<<< [B<-daemons> <I<number of daemons to use>>] >>>
17 S<<< [B<-dcache> <I<number of dcache entries>>] >>> [B<-debug>]
18 [B<-dynroot>] [B<-enable_peer_stats>] [B<-enable_process_stats>]
19 [B<-fakestat>] [B<-fakestat-all>]
20 S<<< [B<-files> <I<files in cache>>] >>>
21 S<<< [B<-files_per_subdir> <I<log(2) of files per dir>> ] >>>
22 [B<-help>] S<<< [B<-logfile> <I<Place to keep the CM log>>] >>>
23 [B<-mem_alloc_sleep>] [B<-memcache>]
24 S<<< [B<-mountdir> <I<mount location>>] >>> [B<-nomount>]
26 S<<< [B<-prealloc> <I<number of 'small' preallocated blocks>>] >>>
27 [B<-rmtsys>] S<<< [B<-rootvol> <I<name of AFS root volume>>] >>>
28 [B<-rxbind>] S<<< [B<-rxmaxmtu> value for maximum MTU ] >>>
29 S<<< [B<-rxpck> value for rx_extraPackets ] >>>
30 [B<-settime>] [B<-shutdown>]
31 S<<< [B<-splitcache> <I<RW/RO ratio>>] >>>
32 S<<< [B<-stat> <I<number of stat entries>>] >>> [B<-verbose>]
33 [B<-disable-dynamic-vcaches>]
34 S<<< [B<-volumes> <I<number of volume entries>>] >>>
42 The B<afsd> command initializes the Cache Manager on an AFS client machine
43 by transferring AFS-related configuration information into kernel memory
44 and starting several daemons. More specifically, the B<afsd> command
45 performs the following actions:
51 Sets a field in kernel memory that defines the machine's cell
52 membership. Some Cache Manager-internal operations and system calls
53 consult this field to learn which cell to execute in. (The AFS command
54 interpreters refer to the F</usr/vice/etc/ThisCell> file instead.) This
55 information is transferred into the kernel from the
56 F</usr/vice/etc/ThisCell> file and cannot be changed until the B<afsd>
61 Places in kernel memory the names and Internet addresses of the database
62 server machines in the local cell and (optionally) foreign cells. The
63 appearance of a cell's database server machines in this list enables the
64 Cache Manager to contact them and to access files in the cell. Omission of
65 a cell from this list, or incorrect information about its database server
66 machines, prevents the Cache Manager from accessing files in it.
68 By default, the list of database server machines is transferred into the
69 kernel from the F</usr/vice/etc/CellServDB> file. Alternatively, when the
70 B<-afsdb> option is used, the list of database server machines is taken
71 from the AFSDB DNS records for each cell. After initialization, use the
72 B<fs newcell> command to change the kernel-resident list without having to
77 Mounts the root of the AFS filespace on a directory on the machine's local
78 disk, according to either the first field in the
79 F</usr/vice/etc/cacheinfo> file (the default) or the B<afsd> command's
80 B<-mountdir> argument. The conventional value is F</afs>.
84 Determines which volume to mount at the root of the AFS file tree. The
85 default is the volume C<root.afs>; use the B<-rootvol> argument to
86 override it. Although the base (read/write) form of the volume name is the
87 appropriate value, the Cache Manager has a bias for accessing the
88 read-only version of the volume (by convention, C<root.afs.readonly>) if
93 Configures the cache on disk (the default) or in machine memory if the
94 B<-memcache> argument is provided. In the latter case, the B<afsd> program
95 allocates space in machine memory for caching, and the Cache Manager uses
96 no disk space for caching even if the machine has a disk.
100 Defines the name of the local disk directory devoted to caching, when the
101 B<-memcache> argument is not used. If necessary, the B<afsd> program
102 creates the directory (its parent directory must already exist). It does
103 not remove the directory that formerly served this function, if one
106 The second field in the F</usr/vice/etc/cacheinfo> file is the source for
107 this name. The standard value is F</usr/vice/cache>. Use the B<-cachedir>
108 argument to override the value in the B<cacheinfo> file.
112 Sets the size of the cache. The default source for the value is the third
113 field in the F</usr/vice/etc/cacheinfo> file, which specifies a number of
116 For a memory cache, the following arguments to the afsd command override
117 the value in the B<cacheinfo> file:
123 The B<-blocks> argument, to specify a different number of kilobyte blocks.
127 The B<-dcache> and B<-chunksize> arguments together, to set both the
128 number of dcache entries and the chunk size (see below for definition of
129 these parameters). In this case, the B<afsd> program derives cache size by
130 multiplying the two values. Using this combination is not recommended, as
131 it requires the issuer to perform the calculation beforehand to determine
132 the resulting cache size.
136 The B<-dcache> argument by itself. In this case, the B<afsd> program
137 derives cache size by multiplying the value specified by the B<-dcache>
138 argument by the default memory cache chunk size of eight kilobytes. Using
139 this argument is not recommended, as it requires the issuer to perform the
140 calculation beforehand to determine the resulting cache size.
144 For satisfactory memory cache performance, the specified value must leave
145 enough memory free to accommodate all other processes and commands that
146 can run on the machine. If the value exceeds the amount of memory
147 available, the B<afsd> program exits without initializing the Cache
148 Manager and produces the following message on the standard output stream:
150 afsd: memCache allocation failure at <number> KB
152 where <number> is how many kilobytes were allocated just before the
155 For a disk cache, use the B<-blocks> argument to the B<afsd> command to
156 override the value in the B<cacheinfo> file. The value specified in either
157 way sets an absolute upper limit on cache size; values provided for other
158 arguments (such as B<-dcache> and B<-chunksize>) never result in a larger
159 cache. The B<afsd> program rejects any setting larger than 95% of the
160 partition size, and exits after generating an error message on the
161 standard output stream, because the cache implementation itself requires a
162 small amount of disk space and overfilling the partition can cause the
163 client machine to panic.
165 To change the size of a disk cache after initialization without rebooting,
166 use the B<fs setcachesize> command; the setting persists until the B<afsd>
167 command runs again or the B<fs setcachesize> command is reissued. The B<fs
168 setcachesize> command does not work for memory caches.
172 Sets the size of each cache I<chunk>, and by implication the amount of
173 data that the Cache Manager requests at a time from the File Server (how
174 much data per fetch RPC, since AFS uses partial file transfer).
176 For a disk cache, a chunk is a F<VI<n>> file and this parameter
177 sets the maximum size to which each one can expand; the default is 64
178 KB. For a memory cache, each chunk is a collection of contiguous memory
179 blocks; the default is size is 8 KB.
181 To override the default chunk size for either type of cache, use the
182 B<-chunksize> argument to provide an integer to be used as an exponent of
183 two; see L<OPTIONS> for details. For a memory cache, if total cache size
184 divided by chunk size leaves a remainder, the B<afsd> program rounds down
185 the number of dcache entries, resulting in a slightly smaller cache.
189 Sets the number of chunks in the cache. For a memory cache, the number of
190 chunks is equal to the cache size divided by the chunk size. For a disk
191 cache, the number of chunks (F<VI<n>> files) is set to the largest
192 of the following unless the B<-files> argument is used to set the value
203 1.5 times the result of dividing cache size by chunk size
204 (I<cachesize>/I<chunksize> * 1.5)
208 The result of dividing cachesize by 10 KB (I<cachesize>/10240)
214 Sets the number of I<dcache entries> allocated in machine memory for
215 storing information about the chunks in the cache.
217 For a disk cache, the F</usr/vice/cache/CacheItems> file contains one
218 entry for each F<VI<n>> file. By default, one half the number of
219 these entries (but not more that 2,000) are duplicated as dcache entries
220 in machine memory for quicker access.
222 For a memory cache, there is no F<CacheItems> file so all information
223 about cache chunks must be in memory as dcache entries. Thus, there is no
224 default number of dcache entries for a memory cache; instead, the B<afsd>
225 program derives it by dividing the cache size by the chunk size.
227 To set the number of dcache entries, use the B<-dcache> argument; the
228 specified value can exceed the default limit of 2,000. Using this argument
229 is not recommended for either type of cache. Increasing the number of
230 dcache entries for a disk cache sometimes improves performance (because
231 more entries are retrieved from memory rather than from disk), but only
232 marginally. Using this argument for a memory cache requires the issuer to
233 calculate the cache size by multiplying this value by the chunk size.
237 Sets the number of I<stat> entries available in machine memory for caching
238 status information about cached AFS files. The default is 300; use the
239 B<-stat> argument to override the default.
243 If the B<-settime> option is specified, then it randomly selects a file
244 server machine in the local cell as the source for the correct time. Every
245 five minutes thereafter, the local clock is adjusted (if necessary) to
246 match the file server machine's clock. This is not enabled by default. It
247 is recommended, instead, that the Network Time Protocol Daemon be used to
248 synchronize the time.
252 In addition to setting cache configuration parameters, the B<afsd> program
253 starts the following daemons. (On most system types, these daemons appear
254 as nameless entries in the output of the UNIX B<ps> command.)
260 One I<callback> daemon, which handles callbacks. It also responds to the
261 File Server's periodic probes, which check that the client machine is
266 One I<maintenance> daemon, which performs the following tasks:
272 Garbage collects obsolete data (for example, expired tokens) from kernel
281 Refreshes information from read-only volumes once per hour.
285 Does delayed writes for NFS clients if the machine is running the NFS/AFS
292 One I<cache-truncation> daemon, which flushes the cache when free space is
293 required, by writing cached data and status information to the File
298 One I<server connection> daemon, which sends a probe to the File
299 Server every few minutes to check that it is still accessible. If the
300 B<-settime> option is set, it also synchronizes the machine's clock
301 with the clock on a randomly-chosen file server machine. There is
302 always one server connection daemon.
306 One or more I<background> daemons that improve performance by pre-fetching
307 files and performing background (delayed) writes of saved data into AFS.
309 The default number of background daemons is two, enough to service at
310 least five simultaneous users of the machine. To increase the number, use
311 the B<-daemons> argument. A value greater than six is not generally
316 On some system types, one I<Rx listener> daemon, which listens for
321 On some system types, one I<Rx event> daemon, which reviews the Rx
322 system's queue of tasks and performs them as appropriate. Most items in
323 the queue are retransmissions of failed packets.
327 On machines that run AIX with virtual memory (VM) integration, one or more
328 I<VM> daemons (sometimes called I<I/O> daemons, which transfer data
329 between disk and machine memory. The number of them depends on the setting
330 of the B<-biods> and B<-daemons> arguments:
336 If the B<-biods> argument is used, it sets the number of VM daemons.
340 If only the B<-daemons> argument is used, the number of VM daemons is
341 twice the number of background daemons.
345 If neither argument is used, there are five VM daemons.
351 This command does not use the syntax conventions of the AFS command
352 suites. Provide the command name and all option names in full.
356 Before using the B<-shutdown> parameter, use the standard UNIX B<umount>
357 command to unmount the AFS root directory (by convention, F</afs>). On
358 Linux, unloading the AFS kernel module and then loading it again before
359 restarting AFS after B<-shutdown> is recommended.
361 AFS has for years had difficulties with being stopped and restarted
362 without an intervening reboot. While most of these issues have been
363 ironed out, stopping and restarting AFS is not recommended unless
364 necessary and rebooting before restarting AFS is still the safest course
365 of action. This does not apply to Linux; it should be safe to restart the
366 AFS client on Linux without rebooting.
368 In contrast to many client-server applications, not all communication is
369 initiated by the client. When the AFS client opens a file, it registers a
370 callback with the AFS server. If the file changes, the server notifies the
371 client that the file has changed and that all cached copies should be
372 discarded. In order to enable full functionality on the AFS client,
373 including all command-line utilities, the following UDP ports must be open
374 on an firewalls between the client and the server:
377 cachemanager 7001/udp (OpenAFS client. Arla uses 4711/udp)
380 kaserver 7004/udp (not needed with Kerberos v5)
382 reserved 7006/udp (for future use)
385 Additionally, for B<klog> to work through the firewall you need to allow
386 inbound and outbound UDP on ports >1024 (probably 1024<port<2048 would
387 suffice depending on the number of simultaneous B<klog>s).
389 Be sure to set the UDP timeouts on the firewall to be at least twenty
390 minutes for the best callback performance.
398 Enable afsdb support. This will use DNS to lookup the AFSDB record and
399 use that for the database servers for each cell instead of the values
400 in the F<CellServDB> file. This has the advantage of only needing to
401 update one DNS record to reconfigure the AFS clients for a new
402 database server as opposed to touching all of the clients, and also
403 allows one to access a cell without preconfiguring its database
404 servers in F<CellServDB>. The format of AFSDB records is defined in
409 Prefer backup volumes for mountpoints in backup volumes. This option means
410 that the AFS client will prefer to resolve mount points to backup volumes
411 when a parent of the current volume is a backup volume. This is similar to
412 the standard behaviour of preferring read-only volumes over read-write
413 volumes when the parent volume is a read-only volume.
415 =item B<-biods> <I<number of I/O daemons>>
417 Sets the number of VM daemons dedicated to performing I/O operations on a
418 machine running a version of AIX with virtual memory (VM) integration. If
419 both this argument and the B<-daemons> argument are omitted, the default
420 is five. If this argument is omitted but the B<-daemons> argument is
421 provided, the number of VM daemons is set to twice the value of the
422 B<-daemons> argument.
424 =item B<-blocks> <I<blocks in cache>>
426 Specifies the number of kilobyte blocks to be made available for caching
427 in the machine's cache directory (for a disk cache) or memory (for a
428 memory cache), overriding the default defined in the third field of the
429 F</usr/vice/etc/cacheinfo> file. For a disk cache, the value cannot exceed
430 95% of the space available in the cache partition. If using a memory
431 cache, do not combine this argument with the B<-dcache> argument, since
432 doing so can possibly result in a chunk size that is not an exponent of 2.
434 =item B<-cachedir> <I<cache directory>>
436 Names the local disk directory to be used as the cache. This value
437 overrides the default defined in the second field of the
438 F</usr/vice/etc/cacheinfo> file.
440 =item B<-chunksize> <I<chunk size>>
442 Sets the size of each cache chunk. The integer provided, which must be
443 from the range C<0> to C<30>, is used as an exponent on the number 2. It
444 overrides the default of 16 for a disk cache (2^16 is 64 KB) and 13 for a
445 memory cache (2^13 is 8 KB). A value of C<0> or less, or greater than
446 C<30>, sets chunk size to the appropriate default. Values less than C<10>
447 (which sets chunk size to a 1 KB) are not recommended. Combining this
448 argument with the B<-dcache> argument is not recommended because it
449 requires that the issuer calculate the cache size that results.
451 B<-chunksize> is an important option when tuning for performance. Setting
452 this option to larger values can increase performance when dealing with
455 =item B<-confdir> <I<configuration directory>>
457 Names a directory other than the F</usr/vice/etc> directory from which to
458 fetch the F<cacheinfo>, F<ThisCell>, and F<CellServDB> configuration
461 =item B<-daemons> <I<number of daemons to use>>
463 Specifies the number of background daemons to run on the machine. These
464 daemons improve efficiency by doing prefetching and background writing of
465 saved data. This value overrides the default of C<2>, which is adequate
466 for a machine serving up to five users. Values greater than C<6> are not
467 generally more effective than C<6>.
469 Note: On AIX machines with integrated virtual memory (VM), the number of
470 VM daemons is set to twice the value of this argument, if it is provided
471 and the B<-biods> argument is not. If both arguments are omitted, there
474 =item B<-dcache> <I<number of dcache entries>>
476 Sets the number of dcache entries in memory, which are used to store
477 information about cache chunks. For a disk cache, this overrides the
478 default, which is 50% of the number of F<VI<n>> files (cache chunks). For
479 a memory cache, this argument effectively sets the number of cache chunks,
480 but its use is not recommended, because it requires the issuer to
481 calculate the resulting total cache size (derived by multiplying this
482 value by the chunk size). Do not combine this argument with the B<-blocks>
483 argument, since doing so can possibly result in a chunk size that is not
488 Generates a highly detailed trace of the B<afsd> program's actions on the
489 standard output stream. The information is useful mostly for debugging
494 The standard behaviour of the AFS client without the B<-dynroot> option is
495 to mount the root.afs volume from the default cell on the F</afs> path. The
496 F</afs> folder and root.afs volume traditionally shows the folders for
497 F<ThisCell> and other cells as configured by the AFS cell administrator.
499 The B<-dynroot> option changes this. Using this option, the AFS client
500 does not mount the root.afs volume on F</afs>. Instead it uses the
501 contents of the F<CellServDB> file to populate the listing of cells in
502 F</afs>. This is known as a DYNamic ROOT. A cell is not contacted until
503 the path F</afs/I<cellname>> if accessed. This functions similarly to an
504 automounter. The main advantage of using B<-dynroot> is that the AFS
505 client will start properly even without network access, whereas the client
506 not using B<-dynroot> will freeze upon startup if cannot contact the
507 default cell specified in F<ThisCell> and mount the root.afs
508 volume. Dynamic root mode is also sometimes called travelling mode because
509 it works well for laptops which don't always have network connectivity.
511 Two advantages of not using dynroot are that listing F</afs> will usually
512 be faster because the contents of F</afs> are limited to what the AFS
513 administrator decides and that symbolic links are traditionally created
514 by the AFS administrator to provide a short name for the cell (i.e.
515 cellname.domain.com is aliased to cellname). However, with dynroot, the
516 local system administrator can limit the default contents of F</afs> by
517 installing a stripped-down F<CellServDB> file, and if dynroot is in effect,
518 the F<CellAlias> file can be used to provide shortname for common AFS cells
519 which provides equivalent functionality to the most commonly used symbolic
522 =item B<-enable_peer_stats>
524 Activates the collection of Rx statistics and allocates memory for their
525 storage. For each connection with a specific UDP port on another machine,
526 a separate record is kept for each type of RPC (FetchFile, GetStatus, and
527 so on) sent or received. To display or otherwise access the records, use
528 the Rx Monitoring API.
530 =item B<-enable_process_stats>
532 Activates the collection of Rx statistics and allocates memory for their
533 storage. A separate record is kept for each type of RPC (FetchFile,
534 GetStatus, and so on) sent or received, aggregated over all connections to
535 other machines. To display or otherwise access the records, use the Rx
540 Return fake values for stat calls on cross-cell mounts. This option makes
541 an C<ls -l> of F</afs> much faster since each cell isn't contacted, and
542 this and the B<-fakestat-all> options are useful on Mac OS X so that the
543 Finder program doesn't try to contact every AFS cell the system knows
546 =item B<-fakestat-all>
548 Return fake values for stat calls on all mounts, not just cross-cell
549 mounts. This and the B<-fakestat> options are useful on Mac OS X so that
550 the Finder program doesn't hang when browsing AFS directories.
552 =item B<-files> <I<files in cache>>
554 Specifies the number of F<VI<n>> files to create in the cache directory
555 for a disk cache, overriding the default that is calculated as described
556 in L<DESCRIPTION>. Each F<VI<n>> file accommodates a chunk of data, and
557 can grow to a maximum size of 64 KB by default. Do not combine this
558 argument with the B<-memcache> argument.
560 =item B<-files_per_subdir> <I<files per cache subdirectory>>
562 Limits the number of cache files in each subdirectory of the cache
563 directory. The value of the option should be the base-two log of the
564 number of cache files per cache subdirectory (so 10 for 1024 files, 14 for
565 16384 files, and so forth).
569 Prints the online help for this command. All other valid options are
572 =item B<-logfile> <I<log file location>>
574 This option is obsolete and no longer has any effect.
576 =item B<-mem_alloc_sleep>
578 Allows sleeps when allocating a memory cache.
582 Initializes a memory cache rather than a disk cache. Do not combine this
583 flag with the B<-files> argument.
585 =item B<-mountdir> <I<mount location>>
587 Names the local disk directory on which to mount the root of the AFS
588 filespace. This value overrides the default defined in the first field of
589 the F</usr/vice/etc/cacheinfo> file. If a value other than the F</afs>
590 directory is used, the machine cannot access the filespace of cells that
595 Do not mount AFS on startup. The afs global mount must be mounted via
596 some other means. This is useful on Mac OS X where /afs is sometimes
597 mounted in /Network/afs like other network file systems.
601 This is enabled by default. It prevents the Cache Manager from
602 synchronizing its clock with the clock on a server machine selected at
603 random by checking the time on the server machine every five minutes.
604 This is the recommended behavior; instead of the AFS Cache Manager, the
605 Network Time Protocol Daemon should be used to synchronize the system
608 =item B<-prealloc> <I<number of preallocated blocks>>
610 Specifies the number of pieces of memory to preallocate for the Cache
611 Manager's internal use. The default initial value is C<400>, but the Cache
612 Manager dynamically allocates more memory as it needs it.
616 Initializes an additional daemon to execute AFS-specific system calls on
617 behalf of NFS client machines. Use this flag only if the machine is an
618 NFS/AFS translator machine serving users of NFS client machines who
619 execute AFS commands.
621 =item B<-rootvol> <I<name of AFS root volume>>
623 Names the read/write volume corresponding to the root directory for the
624 AFS file tree (which is usually the F</afs> directory). This value
625 overrides the default of the C<root.afs> volume. This option is ignored if
626 B<-dynroot> is given.
630 Bind the Rx socket (one interface only).
632 =item B<-rxmaxmtu> <I<value for maximum MTU>>
634 Set a limit for the largest maximum transfer unit (network packet size) that
635 the AFS client on this machine will be willing to transmit. This switch can
636 be used where an artificial limit on the network precludes packets as large
637 as the discoverable MTU from being transmitted successfully.
639 =item B<-rxpck> <I<value for rx_extraPackets>>
641 Set rx_extraPackets to this value. This sets the number of extra Rx
642 packet structures that are available to handle Rx connections. This
643 value should be increased if the "rxdebug 127.0.0.1 -port 7001
644 -rxstats" command shows no free Rx packets. Increasing this value may
645 improve OpenAFS client performance in some circumstances.
649 Enable native AFS time synchronization. This option is the opposite of
650 B<-nosettime> and cannot be used with the B<-nosettime> option.
654 Shuts down the Cache Manager. Before calling B<afsd> with this option,
655 unmount the AFS file system with B<umount>.
657 =item B<-splitcache> <I<RW/RO Ratio>>
659 This allows the user to set a certain percentage of the AFS cache be
660 reserved for read/write content and the rest to be reserved for read-only
661 content. The ratio should be written as a fraction. For example,
662 C<-splitcache 75/25> devotes 75% of your cache space to read/write content
663 and 25% to read-only.
665 =item B<-stat> <I<number of stat entries>>
667 Specifies the number of entries to allocate in the machine's memory for
668 recording status information about the AFS files in the cache. This value
669 overrides the default of C<300>.
673 Generates a detailed trace of the B<afsd> program's actions on the
674 standard output stream.
676 =item B<-volumes> <I<number of volume entries>>
678 Specifies the number of memory structures to allocate for storing volume
679 location information. The default value is C<50>.
681 =item B<-disable-dynamic-vcaches>
683 By default, dynamic vcache overrides the B<-stat> option by using the value of
684 B<-stat> (or the default) as the initial size of the stat (or vcache) pool and
685 increases the pool dynamically as needed on supported platforms. This flag will
686 disable this new functionality and honor the '-stat' setting.
690 Has no effect on the operation of the Cache Manager. The behavior it
691 affected in previous versions of the Cache Manager, to perform synchronous
692 writes to the File Server, is now the default behavior. To perform
693 asynchronous writes in certain cases, use the B<fs storebehind> command.
699 The B<afsd> command is normally included in the machine's AFS
700 initialization file, rather than typed at the command shell prompt. For
701 most disk caches, the appropriate form is
705 The following command is appropriate when enabling a machine to act as an
706 NFS/AFS Translator machine serving more than five users.
708 % /usr/vice/etc/afsd -daemons 4 -rmtsys
710 The following command initializes a memory cache and sets chunk size to 16
713 % /usr/vice/etc/afsd -memcache -chunksize 14
715 =head1 PRIVILEGE REQUIRED
717 The issuer must be logged in as the local superuser root.
726 RFC 1183 L<http://www.faqs.org/rfcs/rfc1183.html>
730 IBM Corporation 2000. <http://www.ibm.com/> All Rights Reserved.
732 This documentation is covered by the IBM Public License Version 1.0. It
733 was converted from HTML to POD by software written by Chas Williams and
734 Russ Allbery, based on work by Alf Wachsmann and Elizabeth Cassell.