Closed mweinelt closed 2 years ago
troglobit
commented
5 years ago Yeah, this seems to be one of the known "bugs", or limitations, in the current implementation. It's even mentioned in the BUGS section in the man page:
pim6sd requires the node running the daemon to have an IPv6 global address.Not sure what to do about that ...
mweinelt
commented
5 years ago ULA is apparently fine, it just wants something other than link-local. But what happened was the following:
default_phyint_status disable;
phyint ip6gre1;
phyint wg-io;This did not enable any interface. After i added phyint [...] enable; it worked, even using a ULA address.
troglobit
commented
5 years ago Aha, good catch! Would you mind writing something up in the README about how to set things up without globally scoped IPv6 address? I think that would really help other people trying pim6sd out.
I'm not even remotely fluent yet in IPv6 so any input from someone else would be greatly appreciated :)
mweinelt
commented
5 years ago Sure thing, let's just wait until we've got things up and running, so we can reduce the number of commits.
troglobit
commented
5 years ago Thanks, but you don't need to worry about that, not for my sake at least. Looking forward to your PR :)
mweinelt
commented
5 years ago As soon as we've exchanged some multicast routes :)
troglobit
commented
5 years ago That's fair. Just remember PIM is quite different from DVMRP (which is flood and prune based). PIM requires a) working unicast routes between peers, and b) someone sending and someone wanting to receive multicast. Otherwise it'll be very quiet. Pretty sure you're aware of this, just a friendly reminder.
mweinelt
commented
5 years ago We, @T-X and I, are playing amongst each other. The idea is to use DN42 (www.dn42.net) as a larger test-bed. Unicast routing is working quite well there, using BGP between AS and a multitude of different IGPs.
troglobit
commented
5 years ago Cool, good luck guys! 😎👍
mweinelt
commented
4 years ago We wrote some docs to get things up and running back then. @T-X just published them here https://dn42.dev/howto/IPv6-Multicast.
troglobit
commented
4 years ago Very nice and informative. Thank you for sharing!
leoplo
commented
2 years ago I do not think I am getting this error for the same reasons as above but it throws the same error in the end. My use case is to run pim6sd in SSM mode with re6st (based on babeld) for unicast routing.
For now I am trying the simple configuration :
default_phyint_status disable;
phyint lo enable;
phyint eth0 enable mld_version any;But this fails :
$ sudo pim6sd -f ssm.conf -n -d all
debug level 0xffffffff: mld_proto, mld_timer, mld_member, mld, switch, trace, mtrace, traceroute, timeout, callout, pkt, packets, interfaces, vif, kernel, cache, mfc, k_cache, k_mfc, rsrr, pim_detail, pim_hello, p
im_neighbors, pim_register, registers, pim_join_prune, pim_j_p, pim_jp, pim_bootstrap, pim_bsr, bsr, bootstrap, pim_asserts, pim_cand_rp, pim_c_rp, pim_rp, rp, pim_routes, pim_routing, pim_mrt, pim_timers, pim_rpf
, rpf, pim, routes, routing, mrt, routers, mrouters, neighbors, timers, asserts, all, 3
17:00:43.044 pim6sd version 2.1.0-beta1 starting
17:00:43.045 65536 octets allocated for the emit/recept buffer mld6
17:00:43.045 Buffer reception size for socket 4 : 262144 in 0 iterations
17:00:43.046 65536 octets allocated for the emit/recept buffer pim6
17:00:43.047 Buffer reception size for socket 5 : 262144 in 0 iterations
17:00:43.047 Interfaces world initialized...
17:00:43.047 Getting vifs from ssm.conf
<3>comment [## SSM configuration example] (28)
<3>comment [## syntax : phyint <interface> <disable> <metric> [metric] <preference> ] (72)
<3>comment [## [preference] [version <version>]] (43)
<3>comment [## metric and pref are for the asserts messages] (49)
<3>comment [#phyint ed1 enable version 1; # speaks only MLDv1] (49)
<3>comment [#phyint de0 enable version 2; # speaks only MLDv2] (49)
<3>comment [#phyint fxp0 enable version any; # speaks MLDv2 and MLDv1-compat mode] (69)
<3>comment [#phyint gif0 disable; # not enables multicast on gif0] (54)
<3>default_phyint_status
<3>disable
<3>end cnf [;] (1)
<3>begin phyint [phyint] (6)
<7>lo
<9>enable
<9>end phyint [;] (1)
<3>begin phyint [phyint] (6)
<7>eth0
<9>enable
<9>mld_version
<9>any
<9>end phyint [;] (1)
<3>comment [## Normally, the following parameters need not be configured.] (61)
<3>comment [##---------------Protocol timer specifications---------------------------#] (74)
<3>comment [## The timer granularity.] (25)
<3>comment [## More this value is small, more pim6sd will be accurate.] (60)
<3>comment [## default if not specified : 5] (33)
<3>comment [## BE SURE to have to same granularity on ALL routers, otherwise....] (70)
<3>comment [#granularity 5;] (15)
<3>comment [## Notes : theses value are the default spec value!] (51)
<3>comment [## do not touch it if you don't know what you do!!] (58)
<3>comment [## you MUST change theses values according to the granularity value!] (76)
<3>comment [## syntax : 'hello_period <number> <coef>'.] (44)
<3>comment [## number is the period in second between 2 hello messages.] (61)
<3>comment [## coef is the coef to deterimine the hello holdtime = hello_period * coef.] (77)
<3>comment [## default if not specified: 30 3.5] (37)
<3>comment [#hello_period 30 3.5;] (21)
<3>comment [## syntax : 'join_prune_period <number> <coef>'.] (49)
<3>comment [## number is the period in second between 2 join/prune messages.] (67)
<3>comment [## coef is the coef to deterimine the join/prune holdtime] (60)
<3>comment [## = join_prune_period * coef.] (33)
<3>comment [## default if not specified : 60 3.5] (36)
<3>comment [#join_prune_period 60 3.5;] (26)
<3>comment [## syntax : 'data_timeout <number>'.] (37)
<3>comment [## number is the time after which (S,G) state for a silent source will ] (74)
<3>comment [## be deleted.] (17)
<3>comment [## default if not specified : 210 ] (38)
<3>comment [#data_timeout 210;] (18)
<3>comment [## syntax : 'assert_timeout <number>'.] (39)
<3>comment [## this is the interval between the last time an Assert is received and ] (75)
<3>comment [## the time at which the assert is timeout.] (46)
<3>comment [## default if not specified : 180] (36)
<3>comment [#assert_timeout 180;] (20)
<3>comment [## syntax : default_source_metric <number>] (42)
<3>comment [#default_source_metric 1243;] (28)
<3>comment [## syntax : default_source_preference <number>] (46)
<3>comment [#default_source_preference 123;] (31)
parse successed.
17:00:43.052 pim_hello_period set to: 30
17:00:43.052 pim_hello_holdtime set to: 105
17:00:43.052 pim_join_prune_period set to: 60
17:00:43.053 pim_join_prune_holdtime set to: 210
17:00:43.053 timer interval set to: 5
17:00:43.053 PIM data timeout set to: 210
17:00:43.053 PIM register suppression timeout set to: 60
17:00:43.053 PIM register probe time set to: 5
17:00:43.054 PIM assert timeout set to: 180
17:00:43.055 mld version for eth0 is v1 v2
17:00:43.055 reg_rate_limit set to 100000 (bits/s)
17:00:43.056 reg_rate_interval set to 20 s.
17:00:43.056 data_rate_limit set to 100000 (bits/s)
17:00:43.056 data_rate_interval set to 20 s.
17:00:43.056 Getting vifs from kernel
17:00:43.058 Installing lo (2401:whatever1 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.058 Installing lo (2401:whatever2 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.059 Installing lo (2401:whatever3 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.059 Installing lo (2401:whatever4 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.059 Installing lo (2401:whatever5 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.059 Installing lo (2401:whatever6 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.060 Installing lo (2401:whatever7 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.060 Installing lo (2401:whatever8 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.060 Installing lo (2401:whatever9 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.061 Installing lo (2401:whatever10 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.061 Installing lo (2401:whatever11 on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.061 Installing lo (::1 on subnet ::/128) ,as vif #2 - rate = 0
17:00:43.062 Installing eth0 (fe80::whatever on subnet ::/64) ,as vif #2 - rate = 0
17:00:43.062 warning - Not registering detected interface re6stnet-tcp: Operation not permitted
17:00:43.062 warning - Not registering detected interface re6stnet1: Operation not permitted
17:00:43.063 warning - Not registering detected interface re6stnet2: Operation not permitted
17:00:43.063 warning - Not registering detected interface re6stnet3: Operation not permitted
17:00:43.063 warning - Not registering detected interface re6stnet4: Operation not permitted
17:00:43.064 warning - Not registering detected interface re6stnet5: Operation not permitted
17:00:43.064 warning - Not registering detected interface re6stnet6: Operation not permitted
17:00:43.065 warning - Not registering detected interface re6stnet7: Operation not permitted
17:00:43.065 warning - Not registering detected interface re6stnet8: Operation not permitted
17:00:43.066 warning - Not registering detected interface re6stnet9: Operation not permitted
17:00:43.066 warning - Not registering detected interface re6stnet10: Operation not permitted
17:00:43.066 There's no global address availableAll my network interfaces except lo have one local ipv6 address and lo has a bunch of global ipv6 addresses.
I do not get why There's no global address available if lo has some. Even if I disable interface eth0 and I keep to Any-source mode (I do not enable MLD), I am getting this error.
I guess the software might not like that global addresses are on the loopback interface.
Did someone else already tried to enable lo interface only ?
troglobit
commented
2 years ago I have to look into this in detail, of course, but could you post the output of ip -br link and ip -br addr? Also, could you try setting up a separate dummy interface instead of using loopback? E.g. ip link add foo0 type dummy; ip link set foo0 multicast on; ...
leoplo
commented
2 years ago olimex@a20-olinuxino:~$ ip -br link
lo UNKNOWN 00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
eth0 UP macAddress1 <BROADCAST,MULTICAST,UP,LOWER_UP>
can0 DOWN <NOARP,ECHO>
enx00e04c88743d DOWN macAddress2 <NO-CARRIER,BROADCAST,MULTICAST,UP>
re6stnet-tcp UNKNOWN macAddress3 <BROADCAST,MULTICAST,UP,LOWER_UP>
re6stnet1 DOWN macAddress4 <BROADCAST,MULTICAST>
re6stnet2 DOWN macAddress5 <BROADCAST,MULTICAST>
re6stnet3 DOWN macAddress6 <BROADCAST,MULTICAST>
re6stnet4 DOWN macAddress7 <BROADCAST,MULTICAST>
re6stnet5 DOWN macAddress8 <BROADCAST,MULTICAST>
re6stnet6 DOWN macAddress9 <BROADCAST,MULTICAST>
re6stnet7 DOWN macAddress10 <BROADCAST,MULTICAST>
re6stnet8 DOWN macAddress11 <BROADCAST,MULTICAST>
re6stnet9 DOWN macAddress12 <BROADCAST,MULTICAST>
re6stnet10 DOWN macAddress13 <BROADCAST,MULTICAST>lo interface not being able to do multicast nor broadcast but having the ipv6 global addresses and the other interfaces (like eth0) being able to do it but only having local ipv6 address could be the issue ?
i.e. Does the current implementation allows to run MLD on an interface with only a local ipv6 address (like eth0 here) and get the global ipv6 from another interface not supporting multicast (lo) ?
user@a20-olinuxino:~$ ip -br addr
lo UNKNOWN 127.0.0.1/8 10.0.whatever1/32 10.0.whatever2/32 10.0.whatever3/32 10.0.whatever4/32 10.0.whatever5/32 10.0.whatever6/32 10.0.whatever7/32 10.0.whatever8/32 10.0.whatever9/32 10.0.whatever10/32 2401:whatever11/64 2401:whatever12/64 2401:whatever13/64 2401:whatever14/64 2401:whatever15/64 2401:whatever16/64 2401:whatever17/64 2401:whatever18/64 2401:whatever19/64 2401:whatever20/64 2401:whatever21/64 ::1/128
eth0 UP 192.168.99.90/24 fe80::321f:whatever/64
can0 DOWN
enx00e04c88743d DOWN 169.0.1.2/8
re6stnet-tcp UNKNOWN fe80::4e8:whatever/64
re6stnet1 DOWN
re6stnet2 DOWN
re6stnet3 DOWN
re6stnet4 DOWN
re6stnet5 DOWN
re6stnet6 DOWN
re6stnet7 DOWN
re6stnet8 DOWN
re6stnet9 DOWN
re6stnet10 DOWN
Pim6sd complains about having no global addresses available, the log shows two (on br-lan and br-wifi).