1 README-3.9-beta3.mrouted,v 1.1.2.1 1998/03/01 03:00:20 fenner Exp
3 IP Multicast Extensions for BSD-Derived Unix Systems
4 Multicast Routing Daemon
9 available from parcftp.xerox.com,
10 file pub/net-research/ipmulti/beta-test/mrouted3.9-beta3.tar.Z
12 pub/net-research/ipmulti/beta-test/mrouted3.9-beta3-sparc-sunos41x.tar.Z
13 pub/net-research/ipmulti/beta-test/mrouted3.9-beta3-sparc-solaris2.tar.Z
14 pub/net-research/ipmulti/beta-test/mrouted3.9-beta3-i386-freebsd22.tar.Z
15 pub/net-research/ipmulti/beta-test/mrouted3.9-beta3-alpha-osf1.tar.Z
16 pub/net-research/ipmulti/beta-test/mrouted3.9-beta3-sgi-irix6.tar.Z
18 Note: The 3.9 release is mrouted-only, and will run on top of a 3.5 kernel.
19 It is a drop-in replacement for mrouted 3.5, 3.6, 3.7 or 3.8.
21 NOTE WELL: This is a beta-test release of mrouted. The basic
22 functionality has been extensively tested in CAIRN and other
23 testbeds, but it is expected to have bugs. Please report bugs to Bill
24 Fenner <fenner@parc.xerox.com>.
27 The 3.9-beta3 release fixes the following bugs:
28 o There was a bug handling routing updates which caused random black
31 o There was a race condition in the timer handlers causing free'd memory
32 to sometimes get touched.
34 o "allow_nonpruners" wasn't allowed in the configuration file (and almost
35 nobody noticed! - probably a good sign)
37 o When a prune times out and the source has been active "recently",
38 mrouted now waits for further traffic instead of triggering a new
41 o mrouted now ignores unreachable routes when making a routing decision
42 (previously it would blackhole, now it can find a less-specific)
44 The 3.9-beta3 release has the following new features:
45 o A "blaster" keyword for mrouted.conf, to turn on handling of routers
46 (mostly ciscos) which overwhelm the socket buffers by blasting the
47 whole routing table at once.
49 o A "notransit" keyword; routes learned on a "notransit" vif will not be
50 readvertised onto another "notransit" vif.
52 o The 500kbps default rate limit on tunnels has been removed.
54 o An ICMP listener which logs ICMP errors which appear to be in response to
55 tunnel packets that we sent.
57 o A tunnel traffic encapsulator, which encapsulates control traffic
58 inside the tunnel instead of unicasting it "beside" the tunnel.
59 This is turned off by default; use "beside off" to turn it on.
61 o A "force_leaf" flag to ignore any potential neighbors on a given interface.
68 The 3.9-beta2 release fixes the following bugs:
69 o There was a bug in 3.9-beta1's raw socket buffer processing that
70 would cause an immediate lockup on startup on some systems.
72 o RSRR would not clear out the group membership information if
73 further notification of changes to this route entry was not possible.
75 There is no need to upgrade to 3.9-beta2 if you are not experiencing
76 one of the aforementioned bugs.
82 The 3.9-beta1 release has the following known bugs:
84 o The startup message doesn't print properly if you have too many
87 The 3.9-beta1 release fixes the following bugs:
89 o mrouted did not properly keep track of subordinates, and would not
90 time out subordinateness. This caused 2 major problems:
91 1. pruning did not happen when there were equal-cost paths to
92 the same multi-access link
93 2. subordinateness which did not get cancelled by a non-poisoned
94 route (e.g. in the face of route filtering) did not time out,
95 causing traffic to continue to flow.
97 o mrouted's IGMPv2 processing when it is not the querier now
98 conforms to draft-ietf-idmr-igmp-v2-06.txt Thanks to Lorenzo
99 VICISANO <L.Vicisano@cs.ucl.ac.uk> for finding a problem.
101 o mrouted is much more careful about forgetting prunes; 3.8
102 would forget prunes whenever any route change ocurred.
104 The 3.9-beta1 release has the following new features:
106 o Longer prune lifetimes (2 hours) by default. Prune lifetimes may
107 be configured per-vif, with the "prune_lifetime N" mrouted.conf
108 configuration file entry (where N is in seconds). This helps to
109 work around the black holes caused on restart when you have a Cisco
110 upstream which does not handle genid's; if this is your situation
111 the recommended value is 300.
113 o mrouted's behavior of flooding new routes by default at startup
114 in order to speed healing of paths during startup can be turned off
115 per-vif or globally with the "noflood" configuration option.
116 Turning this option off means you are likely to experience
117 black holes for a minute or two when you restart a router. The
118 default is to flood for a minute or two until mrouted is able to
119 learn subordinate relationships.
121 o mrouted now retransmits prunes by default on point-to-point links.
122 prune retransmission can be turned on or off per vif via the
123 "rexmit_prunes [on|off]" mrouted.conf command. Prune retransmission
124 helps on lossy links, and also helps when a router has forgotten
125 about a prune (e.g. if it is out of memory and needs to shed state,
128 o The new "passive" mode causes mrouted to not actively send probes
129 looking for neighbors. This allows a dialup link to become quiescent
130 if there is no DVMRP neighbor on the other end. Configuring
131 "passive" on both ends of a link will cause it to never come up.
133 o mrouted defaults to not peering with DVMRP routers that do not
134 prune. Use the "allow_nonpruners" mrouted.conf option on a vif
135 on which you want to allow such peerings.
137 o mrouted now allows route filtering. mrouted.conf syntax:
140 accepts all routes matching 13/8 (e.g. accepts
141 13.2.116/22). If you want to accept only exactly
146 deny 10/8 64/2 130/8 exact 172/8 exact
148 denies some common MBone martians
150 Only "accept" or "deny" is allowed, no combinations.
152 Add "bidir" to apply the filter to output too, otherwise
156 - Providers filter routes that customers send them
158 - Topology modification (e.g. don't let the existence of
159 private tunnel foo out into the world).
162 o mrouted now malloc's the buffer it uses for SIOCGIFCONF, to allow
163 for more interfaces. Thanks to Danny Mitzel
165 o mrouted now ignores multiple entries for a single interface
166 name (temporary hack until mrouted understands interface aliases)
168 o mrouted's "-d" flag has been modified to accept the names of the
169 systems which you would like to debug.
170 packet, prunes, routes, peers, cache, timeout, interface,
171 membership, traceroute, igmp
173 o mrouted now times neighbors out fater, and fully detects and
174 ignores routes from one-way peerings.
176 o mrouted's route processing has been sped up, especially at startup.
178 o mrouted uses the biggest SO_RCVBUF the operating system allows
181 o mrouted uses TOS 0xc0 ("Internet Control") for DVMRP messages.
187 The 3.8 release fixes the following bugs:
189 o mrouted would fail to forget prunes when a neighbor went away,
190 thus potentially sending traffic down a tunnel after the tunnel
191 endpoint has gone down. This was due to some research code making
192 it into the "emergency" 3.7 release, sigh.
194 o mrouted could send prunes with negative lifetimes. This causes
195 slightly higher prune traffic but shouldn't be any major problem.
201 The 3.7 release fixes the following bugs:
203 o mrouted now ignores route reports that include bogus netmasks.
204 There was a bug in 3.5 that would mangle default routes into
205 tens of bogus routes; this should prevent that bug from killing
208 This solution can cause route flaps and black holes until the
209 3.5's are gone or all of the 3.5's neighbors are 3.7 .
211 o mrouted now ignores duplicate routes. Ciscos and the above 3.5
212 bug could cause two copies of the same route to appear in a single
213 routing update; mrouted would insert two copies of the same route
214 into its routing table and wreak all sorts of havoc.
216 o mrouted now sends a group-specific query for both retransmissions
217 of a g-s query; previous versions sent a general query the second
220 o mrouted now loops back multicasted mtrace responses and
221 group-specific membership queries
223 o mrouted now performs deterministic tiebreaking between two
224 neighbors on the same vif.
226 o mrouted now only does duplicate suppression on traceroute requests,
227 not all traceroute packets, so that a loop can be nicely detected
228 via a duplicate router instead of just a timeout.
230 o the buffer size that mrouted uses has been increased to allow
231 more than 16 hops in mtrace messages.
233 o mtrace's hop-by-hop termination is now more likely to be correct.
235 o mrinfo now waits for the responses to its retransmitted queries.
237 The 3.7 release has the following new features:
239 o The configuration file can accept a hostname as the other end
240 of a tunnel. There must be a single name->ip mapping for the
241 given name, however, or mrouted will fail to start up.
243 o mrinfo now sends requests to all interfaces of a multihomed host.
245 o mtrace's passive mode has been implemented.
247 o The first screen of mtrace statistics is shorter and more likely
248 to fit on one screen.
254 The 3.6 release fixes the following bugs:
256 o mrouted would dump core when attempting to report no routes (i.e. upon
257 startup, if you have no enabled phyint's)
259 o mrouted would dump core if requested to traceroute a source for which it
262 o neighbor flags were not always properly updated on probe or report
264 o mrouted would sometimes reply to a multicast traceroute on a disabled
265 phyint; now it uses the first configured phyint to reply to traceroutes.
267 o host routes (i.e. netmask 0xffffffff) works now; it was discarding
268 IGMP from the host because it was coming from the "broadcast address"
271 o send_igmp() now treats the failure to send an mtrace or a neighbor
272 reply as informational, as opposed to warning.
274 o mrouted would go into an infinite loop trying to respond to a traceroute
275 for a source with a netmask of 0xffffffff.
277 o vifs_with_neighbors was not being reset if the mrouted was restarted
280 o the default route was not being properly advertised to neighbors (although
281 it was accepted if it was advertised to it)
283 o ANSI-fication for those who it helps, still-K&R-ish for those it doesn't.
285 o mtrace now attempts to trace three hops past a non-responding router,
286 in the hopes that it does support traceroute but just couldn't respond
287 (i.e. unicast didn't work and it can't source multicast because all its
288 phyints are disabled).
290 o mrinfo now times out even on a multicast router.
297 The 3.5 release has the following new features:
299 o The kernel and mrouted make sure that each is the correct version, to
300 prevent problems with mismatched kernel/mrouted versions. A too-old
301 mrouted will die with the error:
303 can't enable DVMRP routing in kernel: Option not supported by protocol
305 o mrouted can accept and propogate a default route (essential for
306 heirarchical multicast routing)
308 o Kernel route cache keeps source-specific routes instead of subnet routes,
309 eliminating hashing and longest-match problems.
310 (allows classless routing, longest-match and default routing)
312 o Cached kernel routes only get deleted if no traffic is flowing, to
313 facilitate multicast traceroute
315 o mrouted has a new configuration file parser, which provides better error
316 messages than before, and allows named boundaries (see man page)
318 o added "netmask" to phyint configuration, at the suggestion of
321 o System V and FreeBSD compatibility from John Brezak <brezak@ch.hp.com>
323 o phyint's can have additional subnets configured, for people with
324 multiple subnets on one physical network. mrouted.conf syntax is
325 altnet 1.2.3.0, or altnet 1.2.3.0/24 if you need to specify
326 a different netmask. There can be as many altnet statements
329 o both mrouted and the kernel now support classless addresses.
331 o the kernel supports PIM assert processing by notifying the router
332 when a packet arrives on the wrong interface
334 o the kernel keeps additional counters, and mrouted can be compiled to
335 support SNMP and the Multicast MIB
337 o the packet classifier in the kernel now uses the following udp port
339 [0, 16384) - lowest priority, unclassified
340 [16384, 32768) - highest priority, i.e. audio
341 [32768, 49152) - medium priority, i.e. whiteboard
342 [49152, 65536) - low priority, i.e. video
343 A future release of a session directory will allocate ports in these
346 o the configuration code has been modified to default tunnels' rate_limit
347 parameters to 500kbps. This is easily modified with a rate_limit keyword
348 in mrouted.conf, but should be a good default for the MBONE in general.
350 o The tunnel sending code now caches a route for ip_output(), this should
351 help performance on machines with lots of tunnels.
353 o Dispatching for de-capsulating packets is now via protosw[], making
354 reception of other raw protocols more efficient
356 o Neighbor capabilities are discovered via a bitmask as opposed to
359 o Multicast traceroute code improved
361 o mrouted can be compiled with Routing Support for Resource Reservation
362 (RSRR), required for RSVP.
365 The 3.5 release fixes the following bugs:
367 o The IGMPv2 query timeout field was interpreted as being in units of
368 200ms as opposed to 100ms, thus the maximum timeout was set to twice
369 the expected value. This is not fatal, as mrouted always queries
370 twice in the expectation that a packet could get loss, but it does
371 make it less robust in the face of packet loss.
373 o IGMP could report membership in local-only groups (i.e. 224.0.0.X)
375 o IGMP could get confused by hearing its own new membership reports, thus
376 a router would never perform fast leave.
378 o IGMP could reset timers for the wrong interface.
380 o mrouted put a bogus value in the maximum timeout field of IGMPv2 query
383 o Non-querier mrouters would respond to IGMP leave messages
385 o mrouted was not performing fast leave properly
387 o If the last member goes away on a transit network, the upstream router
388 would stop forwarding even if there are downstream members.
390 o Kernel hash function improved
392 o Eliminated possibility of panic(): timeout in cache maintenance
394 o Reordered resource allocation when sending upcall to handle failure properly
396 o some endian-ness bugs squashed in mrouted, probably more to go.
398 o Multicast traceroute could send a reply on a disabled interface.
401 This release consists of the following files:
404 README-3.8.mrouted - this file
406 mrouted/* - version 3.8 of mrouted,
407 mrinfo, map-mbone and
410 ifconfig/* - Changes to ifconfig to
411 show multicast interfaces
413 netstat/* - Diffs to netstat
415 ping/* - sources for ping
419 mtest/* - utility for testing
426 Mrouted 3.8 has two optional features: SNMP and RSRR. RSRR support
427 is required for running RSVP; the SNMP code, with the help of the
428 ISODE snmpd, implements both the Multicast MIB and the DVMRP MIB.
432 Routing Support for Resource Reservations (RSRR) was contributed by
433 Daniel Zappala <daniel@isi.edu>.
435 To enable RSRR support, uncomment the three lines starting with
436 RSRR near the top of the Makefile and "make clean; make". Or use
437 the prebuilt binary, mrouted.rsrr .
439 RSRR allows RSVP to query mrouted for its routing entry for a particular
440 source-group pair. Using the routing entry and the IP_MULTICAST_VIF
441 socket call, RSVP can forward distinct control messages out each
442 outgoing interface. This version of mrouted supports RSRR messages
443 using a Unix datagram socket.
445 RSRR currently includes two pairs of query-reply messages. RSVP sends
446 an Initial Query when it starts. Mrouted responds with an Initial Reply
447 that includes the set of vifs it is using, flagging those that are
448 administratively disabled. When RSVP needs the routing entry for a
449 source-group pair, it sends a Route Query. Mrouted responds with a
450 Route Reply that includes the incoming vif and outgoing vifs for the
453 RSVP may request route change notification by setting the notification
454 bit in the Route Query. If mrouted can provide route change
455 notification for the source-group pair, it sets the notification bit in
456 its Route Reply. When the routing entry for the source-group pair
457 changes, mrouted sends an unsolicited Route Reply containing the new
458 routing information. The initial release of mrouted 3.5 did not support
459 route change notification and always returned a Route Reply with the
460 notification bit cleared. This release of mrouted provides route change
461 notification when possible.
465 SNMP support was contributed by David Thaler <thalerd@eecs.umich.edu>.
467 To enable SNMP support, uncomment the six lines near the top of
468 the Makefile below the description of SNMP support, or use the
469 prebuilt binary, mrouted.snmp or mrouted.rsrr.snmp .
471 To link the SNMP-capable mrouted, you need the CMU libraries.
472 See http://nic.merit.edu/~mbone/ for a full mrouted-snmp distribution.
474 Make sure to add the "sysName", "sysContact", "sysVersion" and
475 "sysLocation" variables to your /etc/mrouted.conf if you want them
476 to provide anything other than default values.
481 sysContact "Bill Fenner <fenner@parc.xerox.com> +1 415 812-4816"
482 sysVersion "SunOS 4.1.3 and mrouted 3.8"
483 sysLocation "MAXC room, PARC building 35"
485 The SNMP version of mrouted has an additional command line flag:
489 Specifies a port for SNMP communication (default 161). This option
490 should be used when another SNMP daemon already exists. The
491 preferred alternate port in this case is port 9161.
493 The mstat(8) program allows querying of statistics using SNMP.