ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/gvpe/doc/gvpe.conf.5.pod
Revision: 1.15
Committed: Mon Dec 5 12:58:08 2005 UTC (18 years, 5 months ago) by pcg
Branch: MAIN
CVS Tags: rel-2_0
Changes since 1.14: +41 -2 lines
Log Message:
*** empty log message ***

File Contents

# User Rev Content
1 pcg 1.1 =head1 NAME
2    
3     gvpe.conf - configuration file for the GNU VPE daemon
4    
5     =head1 SYNOPSIS
6    
7     udp-port = 407
8     mtu = 1492
9     ifname = vpn0
10    
11     node = branch1
12     hostname = 1.2.3.4
13    
14     node = branch2
15     hostname = www.example.net
16     udp-port = 500 # this host uses a different udp-port
17    
18     node = branch3
19     connect = ondemand
20    
21     =head1 DESCRIPTION
22    
23     The gvpe config file consists of a series of lines that contain C<variable
24     = value> pairs. Empty lines are ignored. Comments start with a C<#> and
25     extend to the end of the line. They can be used on their own lines, or
26 pcg 1.13 after any directives. Whitespace is allowed around the C<=> sign or after
27     values, but not within the variable names or values themselves.
28 pcg 1.1
29     The only exception to the above is the "on" directive that can prefix any
30     C<name = value> setting and will only "execute" it on the named node, or
31     (if the nodename starts with "!") on all nodes except the named one.
32    
33     name = value
34     on branch1 loglevel = noise
35     on !branch2 connect = ondemand
36    
37     All settings are executed "in order", that is, later settings of the same
38     variable overwrite earlier ones.
39    
40     =head1 ANATOMY OF A CONFIG FILE
41    
42     Usually, a config file starts with global settings (like the udp port to
43     listen on), followed by node-specific sections that begin with a C<node =
44     nickname> line.
45    
46     Every node that is part of the network must have a section that starts
47     with C<node = nickname>. The number and order of the nodes is important
48     and must be the same on all hosts. It is not uncommon for node sections to
49     be completely empty - if the default values are right.
50    
51     Node-specific settings can be used at any time. If used before the first
52     node section they will set the default values for all following nodes.
53    
54     =head1 CONFIG VARIABLES
55    
56     =head2 GLOBAL SETTINGS
57    
58     Global settings will affect the behaviour of the running gvpe daemon, that
59     is, they are in some sense node-specific (config files can set different
60     values on different nodes using C<on>), but will affect the behaviour of
61     the gvpe daemon and all connections it creates.
62    
63     =over 4
64    
65 pcg 1.6 =item dns-forw-host = hostname/ip
66 pcg 1.1
67 pcg 1.6 The dns server to forward dns requests to for the DNS tunnel protocol
68     (default: C<127.0.0.1>, changing it is highly recommended).
69 pcg 1.1
70 pcg 1.6 =item dns-forw-port = port-number
71 pcg 1.1
72 pcg 1.6 The port where the C<dns-forw-host> is to be contacted (default: C<53>,
73     which is fine in most cases).
74 pcg 1.1
75 pcg 1.12 =item dns-max-outstanding = integer-number-of-requests
76    
77     The maximum number of outstanding DNS transport requests
78     (default: C<100>). GVPE will never issue more requests then the given
79     limit without receiving replies. In heavily overloaded situations it might
80     help to set this to a low number (e.g. C<3> or even C<1>) to limit the
81     number of parallel requests.
82    
83     The default should be working ok for most links.
84    
85     =item dns-overlap-factor = float
86    
87     The DNS transport uses the minimum request latency (B<min_latency>) seen
88     during a connection as it's timing base. This factor (default: C<0.5>,
89     must be > 0) is multiplied by B<min_latency> to get the maximum sending
90     rate (= minimum send interval), i.e. a factor of C<1> means that a new
91     request might be generated every B<min_latency> seconds, which means on
92     average there should only ever be one outstanding request. A factor of
93     C<0.5> means that GVPE will send requests twice as often as the minimum
94     latency measured.
95    
96     For congested or picky dns forwarders you could use a value nearer to or
97     exceeding C<1>.
98    
99     The default should be working ok for most links.
100    
101     =item dns-send-interval = send-interval-in-seconds
102    
103     The minimum send interval (= maximum rate) that the DNS transport will
104     use to send new DNS requests. GVPE will not exceed this rate even when
105     the latency is very low. The default is C<0.01>, which means GVPE will
106     not send more than 100 DNS requests per connection per second. For
107     high-bandwidth links you could go lower, e.g. to C<0.001> or so. For
108     congested or rate-limited links, you might want to go higher, say C<0.1>,
109     C<0.2> or even higher.
110    
111     The default should be working ok for most links.
112    
113     =item dns-timeout-factor = float
114    
115     Factor to multiply the C<min_latency> (see C<dns-overlap-factor>) by to
116     get request timeouts. The default of C<8> means that the DNS transport
117     will resend the request when no reply has been received for longer than
118     eight times the minimum (= expected) latency, assuming the request or
119     reply has been lost.
120    
121     For congested links a higher value might be necessary (e.g. C<30>). If the
122     link is very stable lower values (e.g. C<2>) might work nicely. Values
123     near or below C<1> makes no sense whatsoever.
124    
125     The default should be working ok for most links.
126    
127 pcg 1.1 =item if-up = relative-or-absolute-path
128    
129     Sets the path of a script that should be called immediately after the
130     network interface is initialized (but not neccessarily up). The following
131 pcg 1.13 environment variables are passed to it (the values are just examples).
132    
133     Variables that have the same value on all nodes:
134 pcg 1.1
135     =over 4
136    
137     =item CONFBASE=/etc/gvpe
138    
139     The configuration base directory.
140    
141     =item IFNAME=vpn0
142    
143 pcg 1.13 The network interface to initialize.
144    
145     =item IFTYPE=native # or tincd
146    
147     =item IFSUBTYPE=linux # or freebsd, darwin etc..
148    
149     The interface type (C<native> or C<tincd>) and the subtype (usually the
150     OS name in lowercase) that this GVPE was configured for. Can be used to
151     select the correct syntax to use for network-related commands.
152 pcg 1.1
153     =item MTU=1436
154    
155     The MTU to set the interface to. You can use lower values (if done
156     consistently on all hosts), but this is usually ineffective.
157    
158 pcg 1.13 =item NODES=5
159    
160     The number of nodes in this GVPE network.
161    
162     =back
163    
164     Variables that are node-specific and with values pertaining to the node
165     running this GVPE:
166 pcg 1.1
167 pcg 1.13 =over 4
168 pcg 1.1
169 pcg 1.13 =item IFUPDATA=string
170 pcg 1.1
171 pcg 1.13 The value of the configuration directive C<if-up-data>.
172 pcg 1.1
173 pcg 1.13 =item MAC=fe:fd:80:00:00:01
174 pcg 1.1
175 pcg 1.13 The MAC address the network interface has to use.
176 pcg 1.1
177 pcg 1.13 Might be used to initialize interfaces on platforms where GVPE does not
178     do this automatically. Please see the C<gvpe.osdep(5)> manpage for
179     platform-specific information.
180 pcg 1.1
181     =item NODENAME=branch1
182    
183 pcg 1.13 The nickname of the node.
184 pcg 1.1
185     =item NODEID=1
186    
187 pcg 1.13 The numerical node ID of the node running this instance of GVPE. The first
188     node mentioned in the config file gets ID 1, the second ID 2 and so on.
189 pcg 1.1
190     =back
191    
192 pcg 1.13 In addition, all node-specific variables (except C<NODEID>) will be
193     available with a postfix of C<_nodeid>, which contains the value for that
194     node, e.g. the C<MAC_1> variable contains the MAC address of node #1, while
195     the C<NODENAME_22> variable contains the name of node #22.
196    
197 pcg 1.1 Here is a simple if-up script:
198    
199     #!/bin/sh
200 pcg 1.13 ip link set $IFNAME up
201 pcg 1.1 [ $NODENAME = branch1 ] && ip addr add 10.0.0.1 dev $IFNAME
202     [ $NODENAME = branch2 ] && ip addr add 10.1.0.1 dev $IFNAME
203     ip route add 10.0.0.0/8 dev $IFNAME
204    
205     More complicated examples (using routing to reduce arp traffic) can be
206     found in the etc/ subdirectory of the distribution.
207    
208 pcg 1.6 =item ifname = devname
209    
210     Sets the tun interface name to the given name. The default is OS-specific
211     and most probably something like C<tun0>.
212    
213     =item ifpersist = yes|true|on | no|false|off
214    
215     Should the tun/tap device be made persistent, that is, should the device
216     stay up even when gvpe exits? Some versions of the tunnel device have
217     problems sending packets when gvpe is restarted in persistent mode, so
218     if the connections can be established but you cannot send packets from
219     the local node, try to set this to C<off> and do an ifconfig down on the
220     device.
221    
222     =item ip-proto = numerical-ip-protocol
223    
224     Sets the protocol number to be used for the rawip protocol. This is a
225     global option because all hosts must use the same protocol, and since
226     there are no port numbers, you cannot easily run more than one gvpe
227     instance using the same protocol, nor can you share the protocol with
228     other programs.
229    
230     The default is 47 (GRE), which has a good chance of tunneling through
231     firewalls (but note that the rawip protocol is not GRE compatible). Other
232     common choices are 50 (IPSEC, ESP), 51 (IPSEC, AH), 4 (IPIP tunnels) or 98
233     (ENCAP, rfc1241)
234    
235     =item http-proxy-host = hostname/ip
236    
237     The C<http-proxy-*> family of options are only available if gvpe was
238     compiled with the C<--enable-http-proxy> option and enable tunneling of
239     tcp connections through a http proxy server.
240    
241     C<http-proxy-host> and C<http-proxy-port> should specify the hostname and
242     port number of the proxy server. See C<http-proxy-loginpw> if your proxy
243     requires authentication.
244    
245     Please note that gvpe will still try to resolve all hostnames in the
246     configuration file, so if you are behind a proxy without access to a dns
247     server better use numerical IP addresses.
248    
249     To make best use of this option disable all protocols except tcp in your
250     config file and make sure your routers (or all other hosts) are listening
251     on a port that the proxy allows (443, https, is a common choice).
252    
253     If you have a router, connecting to it will suffice. Otherwise tcp must be
254     enabled on all hosts.
255    
256     Example:
257    
258     http-proxy-host = proxy.example.com
259     http-proxy-port = 3128 # 8080 is another common choice
260     http-proxy-auth = schmorp:grumbeere
261    
262     =item http-proxy-port = proxy-tcp-port
263    
264     The port where your proxy server listens.
265    
266     =item http-proxy-auth = login:password
267    
268     The optional login and password used to authenticate to the proxy server,
269     seperated by a literal colon (C<:>). Only basic authentication is
270     currently supported.
271    
272     =item keepalive = seconds
273    
274     Sets the keepalive probe interval in seconds (default: C<60>). After this
275     many seconds of inactivity the daemon will start to send keepalive probe
276     every 5 seconds until it receives a reply from the other end. If no reply
277     is received within 30 seconds, the peer is considered unreachable and the
278     connection is closed.
279    
280     =item loglevel = noise|trace|debug|info|notice|warn|error|critical
281    
282     Set the logging level. Connection established messages are logged at level
283     C<info>, notable errors are logged with C<error>. Default is C<info>.
284    
285     =item mtu = bytes
286    
287     Sets the maximum MTU that should be used on outgoing packets (basically
288     the MTU of the outgoing interface) The daemon will automatically calculate
289     maximum overhead (e.g. udp header size, encryption blocksize...) and pass
290     this information to the C<if-up> script.
291    
292     Recommended values are 1500 (ethernet), 1492 (pppoe), 1472 (pptp).
293    
294     This value must be the minimum of the mtu values of all hosts.
295    
296     =item node = nickname
297    
298     Not really a config setting but introduces a node section. The nickname is
299     used to select the right configuration section and must be passed as an
300     argument to the gvpe daemon.
301    
302 pcg 1.1 =item node-up = relative-or-absolute-path
303    
304     Sets a command (default: no script) that should be called whenever a
305 pcg 1.13 connection is established (even on rekeying operations). In addition to
306     all the variables passed to C<if-up> scripts, the following environment
307 pcg 1.1 variables will be set:
308    
309     =over 4
310    
311     =item DESTNODE=branch2
312    
313     The name of the remote node.
314    
315     =item DESTID=2
316    
317     The node id of the remote node.
318    
319     =item DESTIP=188.13.66.8
320    
321     The numerical IP address of the remote host (gvpe accepts connections from
322     everywhere, as long as the other host can authenticate itself).
323    
324     =item DESTPORT=655 # deprecated
325    
326     The UDP port used by the other side.
327    
328     =item STATE=UP
329    
330     Node-up scripts get called with STATE=UP, node-down scripts get called
331     with STATE=DOWN.
332    
333     =back
334    
335     Here is a nontrivial example that uses nsupdate to update the name => ip
336     mapping in some dns zone:
337    
338     #!/bin/sh
339     {
340     echo update delete $DESTNODE.lowttl.example.net. a
341     echo update add $DESTNODE.lowttl.example.net. 1 in a $DESTIP
342     echo
343     } | nsupdate -d -k $CONFBASE:key.example.net.
344    
345     =item node-down = relative-or-absolute-path
346    
347     Same as C<node-up>, but gets called whenever a connection is lost.
348    
349 pcg 1.6 =item pid-file = path
350    
351     The path to the pid file to check and create
352     (default: C<LOCALSTATEDIR/run/gvpe.pid>).
353    
354     =item private-key = relative-path-to-key
355    
356     Sets the path (relative to the config directory) to the private key
357     (default: C<hostkey>). This is a printf format string so every C<%> must
358     be doubled. A single C<%s> is replaced by the hostname, so you could
359     use paths like C<hostkeys/%s> to fetch the files at the location where
360     C<gvpectrl> puts them.
361 pcg 1.1
362 pcg 1.6 Since only the private key file of the current node is used and the
363     private key file should be kept secret per-host to avoid spoofings, it is
364     not recommended to use this feature.
365 pcg 1.1
366 pcg 1.6 =item rekey = seconds
367 pcg 1.1
368 pcg 1.6 Sets the rekeying interval in seconds (default: C<3600>). Connections are
369     reestablished every C<rekey> seconds.
370 pcg 1.1
371 pcg 1.6 =back
372 pcg 1.1
373 pcg 1.6 =head2 NODE SPECIFIC SETTINGS
374 pcg 1.1
375 pcg 1.6 The following settings are node-specific, that is, every node can have
376     different settings, even within the same gvpe instance. Settings that are
377 pcg 1.15 set before the first node section set the defaults, settings that are
378     set within a node section only apply to the given node.
379 pcg 1.1
380 pcg 1.6 =over 4
381 pcg 1.1
382 pcg 1.15 =item allow-direct = nodename
383    
384     Allow direct connections to this node. See C<deny-direct> for more info.
385    
386 pcg 1.6 =item compress = yes|true|on | no|false|off
387 pcg 1.1
388 pcg 1.6 Wether to compress data packets sent to this host (default: C<yes>).
389     Compression is really cheap even on slow computers and has no size
390     overhead at all, so enabling this is a good idea.
391 pcg 1.1
392 pcg 1.6 =item connect = ondemand | never | always | disabled
393 pcg 1.1
394 pcg 1.6 Sets the connect mode (default: C<always>). It can be C<always> (always
395     try to establish and keep a connection to the given host), C<never>
396     (never initiate a connection to the given host, but accept connections),
397     C<ondemand> (try to establish a connection on the first packet sent, and
398     take it down after the keepalive interval) or C<disabled> (node is bad,
399     don't talk to it).
400 pcg 1.1
401 pcg 1.15 =item deny-direct = nodename | *
402    
403     Deny direct connections to the specified node (or all nodes when C<*>
404     is given). Only one node can be specified, but you can use multiple
405     C<allow-direct> and C<deny-direct> statements. This only makes sense in
406     networks with routers, as routers are required for indirect connections.
407    
408     Sometimes, a node cannot reach some other nodes for reasons of network
409     connectivity. For example, a node behind a firewall that only allows
410     conenctions to/from a single other node in the network. In this case one
411     should specify C<deny-direct = *> and C<allow-direct = othernodename> (the other
412     node I<must> be a router for this to work).
413    
414     The algorithm to check wether a connection may be direct is as follows:
415    
416     1. Other node mentioned in a C<allow-direct>? If yes, allow the connection.
417    
418     2. Other node mentioned in a C<deny-direct>? If yes, deny direct connections.
419    
420     3. Allow the connection.
421    
422     That is, C<allow-direct> takes precende over C<deny-direct>.
423    
424     The check is done in both directions, i.e. both nodes must allow a direct
425     connection before one is attempted, so you only need to specify connect
426     limitations on one node.
427    
428 pcg 1.6 =item dns-domain = domain-suffix
429 pcg 1.1
430 pcg 1.7 The DNS domain suffix that points to the DNS tunnel server for this node.
431 pcg 1.1
432 pcg 1.6 The domain must point to a NS record that points to the I<dns-hostname>,
433     i.e.
434 pcg 1.1
435 pcg 1.6 dns-domainname = tunnel.example.net
436     dns-hostname = tunnel-server.example.net
437 pcg 1.1
438 pcg 1.6 Corresponds to the following DNS entries in the C<example.net> domain:
439 pcg 1.1
440 pcg 1.6 tunnel.example.net. NS tunnel-server.example.net.
441     tunnel-server.example.net. A 13.13.13.13
442 pcg 1.1
443 pcg 1.6 =item dns-hostname = hostname/ip
444 pcg 1.1
445 pcg 1.6 The address to bind the DNS tunnel socket to, similar to the C<hostname>,
446     but for the DNS tunnel protocol only. Default: C<0.0.0.0>, but that might
447     change.
448 pcg 1.1
449 pcg 1.6 =item dns-port = port-number
450 pcg 1.1
451 pcg 1.8 The port to bind the DNS tunnel socket to. Must be C<53> on DNS tunnel servers.
452 pcg 1.1
453 pcg 1.7 =item enable-dns = yes|true|on | no|false|off
454    
455 pcg 1.10 See gvpe.protocol(7) for a description of the DNS transport
456     protocol. Avoid this protocol if you can.
457    
458 pcg 1.8 Enable the DNS tunneling protocol on this node, either as server or as
459 pcg 1.10 client. Support for this transport protocol is only available when gvpe
460     was compiled using the C<--enable-dns> option.
461    
462     =item enable-icmp = yes|true|on | no|false|off
463    
464     See gvpe.protocol(7) for a description of the ICMP transport protocol.
465 pcg 1.8
466 pcg 1.10 Enable the ICMP transport using icmp packets of type C<icmp-type> on this
467     node.
468 pcg 1.7
469 pcg 1.1 =item enable-rawip = yes|true|on | no|false|off
470    
471 pcg 1.10 See gvpe.protocol(7) for a description of the RAW IP transport protocol.
472    
473 pcg 1.1 Enable the RAW IPv4 transport using the C<ip-proto> protocol
474 pcg 1.10 (default: C<no>).
475 pcg 1.1
476 pcg 1.6 =item enable-tcp = yes|true|on | no|false|off
477    
478 pcg 1.10 See gvpe.protocol(7) for a description of the TCP transport protocol.
479    
480 pcg 1.6 Enable the TCPv4 transport using the C<tcp-port> port
481 pcg 1.10 (default: C<no>). Support for this transport protocol is only available
482     when gvpe was compiled using the C<--enable-tcp> option.
483 pcg 1.6
484 pcg 1.1 =item enable-udp = yes|true|on | no|false|off
485    
486 pcg 1.10 See gvpe.protocol(7) for a description of the UDP transport protocol.
487    
488 pcg 1.5 Enable the UDPv4 transport using the C<udp-port> port (default: C<no>,
489     unless no other protocol is enabled for a node, in which case this
490 pcg 1.10 protocol is enabled automatically).
491 pcg 1.5
492     NOTE: Please specify C<enable-udp = yes> if you want t use it even though
493     it might get switched on automatically, as some future version might
494     default to another default protocol.
495 pcg 1.15
496     =item hostname = hostname | ip [can not be defaulted]
497    
498     Forces the address of this node to be set to the given dns hostname or ip
499     address. It will be resolved before each connect request, so dyndns should
500     work fine. If this setting is not specified and a router is available,
501     then the router will be queried for the address of this node. Otherwise,
502     the connection attempt will fail.
503 pcg 1.1
504 pcg 1.11 =item icmp-type = integer
505    
506     Sets the type value to be used for outgoing (and incoming) packets sent
507     via the ICMP transport.
508    
509     The default is C<0> (which is C<echo-reply>, also known as
510     "ping-replies"). Other useful values include C<8> (C<echo-request>, a.k.a.
511     "ping") and C<11> (C<time-exceeded>), but any 8-bit value can be used.
512    
513 pcg 1.13 =item if-up-data = value
514    
515     The value specified using this directive will be passed to the C<if-up>
516     script in the environment variable C<IFUPDATA>.
517    
518 pcg 1.6 =item inherit-tos = yes|true|on | no|false|off
519    
520     Wether to inherit the TOS settings of packets sent to the tunnel when
521     sending packets to this node (default: C<yes>). If set to C<yes> then
522     outgoing tunnel packets will have the same TOS setting as the packets sent
523     to the tunnel device, which is usually what you want.
524    
525     =item max-retry = positive-number
526 pcg 1.1
527 pcg 1.8 The maximum interval in seconds (default: C<3600>, one hour) between
528 pcg 1.6 retries to establish a connection to this node. When a connection cannot
529     be established, gvpe uses exponential backoff capped at this value. It's
530     sometimes useful to set this to a much lower value (e.g. C<120>) on
531     connections to routers that usually are stable but sometimes are down, to
532 pcg 1.8 assure quick reconnections even after longer downtimes.
533 pcg 1.1
534 pcg 1.8 =item router-priority = 0 | 1 | positive-number>=2
535 pcg 1.1
536     Sets the router priority of the given host (default: C<0>, disabled). If
537     some host tries to connect to another host without a hostname, it asks
538     the router host for it's IP address. The router host is the one with the
539 pcg 1.2 highest priority larger than C<1> that is currently reachable.
540 pcg 1.1
541 pcg 1.2 Make sure all hosts always connect (C<connect = always>) to the router
542     hosts, otherwise connecting to them might be impossible.
543    
544     The special value C<1> allows other hosts to route through the router
545     host, but they will never route through it by default. The value C<0>
546     disables routing. The idea behind this is that some hosts can, if
547     required, bump the C<router-priority> setting to higher than C<1> in their
548     local config to route through specific hosts. If C<router-priority> is
549     C<0>, then routing will be refused, so C<1> serves as a "enable, but do
550     not use by default" switch.
551    
552 pcg 1.6 =item tcp-port = port-number
553 pcg 1.1
554 pcg 1.6 Similar to C<udp-port> (default: C<655>), but sets the TCP port number.
555 pcg 1.1
556 pcg 1.6 =item udp-port = port-number
557 pcg 1.1
558 pcg 1.6 Sets the port number used by the UDP protocol (default: C<655>, not
559     officially assigned by IANA!).
560 pcg 1.1
561     =back
562    
563     =head1 CONFIG DIRECTORY LAYOUT
564    
565     The default (or recommended) directory layout for the config directory is:
566    
567     =over 4
568    
569 pcg 1.4 =item X<gvpe.conf>
570 pcg 1.1
571     The config file.
572    
573 pcg 1.4 =item X<if-up>
574 pcg 1.1
575     The if-up script
576    
577 pcg 1.4 =item X<node-up>, X<node-down>
578 pcg 1.1
579     If used the node up or node-down scripts.
580    
581 pcg 1.4 =item X<hostkey>
582 pcg 1.1
583     The private key (taken from C<hostkeys/nodename>) of the current host.
584    
585 pcg 1.4 =item X<pubkey/nodename>
586 pcg 1.1
587     The public keys of the other nodes, one file per node.
588    
589     =back
590    
591     =head1 SEE ALSO
592    
593     gvpe(5), gvpe(8), gvpectrl(8).
594    
595     =head1 AUTHOR
596    
597 pcg 1.14 Marc Lehmann <gvpe@schmorp.de>
598 pcg 1.1