April 2019 - Bird-users - Mailing Lists

[PATCH] Nest: "Last route change" timestamp for routing tables
by Tom Petty 30 Apr '19

30 Apr '19

I¹ve written a patch which adds a ³last route change² timestamp to the routing table statistics. This diff is against the Œmaarten¹ branch, where the code adding the ³show table statistics² command is currently located. diff --git a/nest/route.h b/nest/route.h index c7c44ecc..8472b432 100644 --- a/nest/route.h +++ b/nest/route.h @@ -166,6 +166,7 @@ typedef struct rtable { uint route_count; /* Number of routes in the table */ uint route_updates; /* Number of accepted route updates */ uint route_withdraws; /* Number of accepted route withdraws */ + btime last_route_change; /* Time of last route change */ struct hostcache *hostcache; struct rtable_config *config; /* Configuration of this table */ struct config *deleted; /* Table doesn't exist in current configuration, diff --git a/nest/rt-table.c b/nest/rt-table.c index efc1f1d7..116e8756 100644 --- a/nest/rt-table.c +++ b/nest/rt-table.c @@ -1143,9 +1143,9 @@ rte_recalculate(struct channel *c, net *net, rte *new, struct rte_src *src) skip_stats1: if (new_ok) - table->route_count++; + { table->route_count++; table->last_route_change = current_time(); } if (old_ok) - table->route_count--; + { table->route_count--; table->last_route_change = current_time(); } if (new) rte_is_filtered(new) ? stats->filt_routes++ : stats->imp_routes++; @@ -1668,6 +1668,7 @@ rt_setup(pool *p, rtable *t, struct rtable_config *cf) t->rt_event->hook = rt_event; t->rt_event->data = t; t->gc_time = current_time(); + t->last_route_change = current_time(); } /** @@ -2299,7 +2300,7 @@ rte_update_in(struct channel *c, const net_addr *n, rte *new, struct rte_src *sr *pos = old->next; rte_free_quick(old); tab->route_count--; - + tab->last_route_change = current_time(); break; } @@ -2333,6 +2334,7 @@ rte_update_in(struct channel *c, const net_addr *n, rte *new, struct rte_src *sr e->next = *pos; *pos = e; tab->route_count++; + tab->last_route_change = current_time(); return 1; drop_update: @@ -2406,6 +2408,7 @@ rt_prune_sync(rtable *t, int all) *ee = e->next; rte_free_quick(e); t->route_count--; + t->last_route_change = current_time(); } else ee = &e->next; @@ -2711,6 +2714,9 @@ rt_get_hostentry(rtable *tab, ip_addr a, ip_addr ll, rtable *dep) static void rt_show_table_stats(rtable *tab) { + byte tm[TM_DATETIME_BUFFER_SIZE]; + tm_format_time(tm, &config->tf_route, tab->last_route_change); + cli_msg(-1026, "%s:", tab->name); cli_msg(-1026, "Table type:\t\t%s", net_label[tab->addr_type]); cli_msg(-1026, "Hash table size:\t%u", tab->fib.hash_size); @@ -2718,6 +2724,7 @@ rt_show_table_stats(rtable *tab) cli_msg(-1026, "Total route count:\t%u", tab->route_count); cli_msg(-1026, "Route updates:\t\t%u", tab->route_updates); cli_msg(-1026, "Route withdraws:\t%u", tab->route_withdraws); + cli_msg(-1026, "Last route update:\t%s", tm); cli_msg(-1026, ""); cli_msg(-1026, "%-16s %10s %10s %10s %10s", "Protocol", "Routes", "Preferred", "Updates", "Withdraws"); Thanks, Tom P

2 2

RFC8203bis
by Alexander Azimov 30 Apr '19

30 Apr '19

Hi all, I've just submitted a two-line patch to make BIRD compliant with RFC8203bis <https://datatracker.ietf.org/doc/draft-ietf-idr-rfc8203bis/>: https://github.com/BIRD/bird/pull/5 Please let me know if you have any questions. -- Best regards, Alexander Azimov

2 1

Static dormant routes not shown
by Kenth Eriksson 29 Apr '19

29 Apr '19

Hi! Why does not static routes in dormant state show of up in the routing table? I expexted all routes to be visible in 'show route'. If the route is not installed into the kernel, then that should be some way (in bird case I believe that is the exclamation mark). Thanks, Kenth

2 2

RFC7772 Unicast RAdv Replies
by Martin Weinelt 29 Apr '19

29 Apr '19

Hi all, we're users of Birds RAdv implementation in a mesh networking project with many opportunistic WiFi users on mobile devices. We and our users would likely profit from a reduced rate of router advertisments that we see as a result from high client turnover. RFC7772¹ describes in 5.1: Router manufacturers SHOULD allow network administrators to configure the routers to respond to Router Solicitations with unicast Router Advertisements if: * The Router Solicitation's source address is not the unspecified address, and: * The solicitation contains a valid Source Link-Layer Address option. 2. Administrators of networks that serve large numbers (tens or hundreds) of battery-powered devices SHOULD enable this behavior. 3. Networks that serve battery-powered devices SHOULD NOT send multicast RAs too frequently (see Section 4) unless the information in the RA packet has substantially changed. If there is a desire to ensure that hosts pick up configuration changes quickly, those networks MAY send frequent Router Advertisements for a limited period of time (e.g., not more than one minute) immediately after a configuration change. There is already an implementation for this behaviour in radvd² with the peculiarity that they ignore missing SLLA, because little to no clients actually sent them back in 2017. Note that the RFC does not make SLLA a MUST. They justify unicast responses in absence of SLLA as follows: However, testing shows that many clients do not set the SLLA option: https://github.com/reubenhwk/radvd/issues/63#issuecomment-287172252 As of 2017/03/16: - macOS 10.12.3 sierra - sends SLLA 2 times out of 4 - iOS 10.2.1 (iPhone 5s) - no SLLA - Android 7.0 (sony xperia phone) - sends SLLA - Android 5.1 (nexus 7 tablet) - sends SLLA - Ubuntu 16.04.2 LTS w/ Network Manager, running 4.9 kernel (dell laptop) - no SLLA - Windows 10 (dell laptop) - no SLLA We decide to ignore the SLLA option for now, and only require the unspecified address check. Clients that did not set the SLLA option will trigger a neighbour solicit to the solicited-node address trying to resolve the link-local address to, this would still be less traffic than the all-nodes multicast. I'd be very happy if you would consider implementing this BCP and would especially appreciate it if there were a knob to make Bird reply with unicast radv replies to router solications and only use multicast advertisments in larger intervals. Thanks for your consideration! Martin Weinelt [1] https://tools.ietf.org/html/rfc7772 [2] https://github.com/reubenhwk/radvd/commit/30328bdbfc1f08ea100de9a6011f981cc…

2 1

[PATCH] BGP: Output Local AS number in show protocol
by Jan-Philipp Litza 29 Apr '19

29 Apr '19

This eases the implementation of agents implementing the SNMP-BGP MIB, which requires the local AS of a session to be specified. --- proto/bgp/bgp.c | 1 + 1 file changed, 1 insertion(+) diff --git a/proto/bgp/bgp.c b/proto/bgp/bgp.c index c332a836..9013e8ea 100644 --- a/proto/bgp/bgp.c +++ b/proto/bgp/bgp.c @@ -2259,6 +2259,7 @@ bgp_show_proto_info(struct proto *P) cli_msg(-1006, " BGP state: %s", bgp_state_dsc(p)); cli_msg(-1006, " Neighbor address: %I%J", p->cf->remote_ip, p->cf->iface); cli_msg(-1006, " Neighbor AS: %u", p->cf->remote_as ?: p->remote_as); + cli_msg(-1006, " Local AS: %u", p->cf->local_as ?: p->local_as); if (p->gr_active_num) cli_msg(-1006, " Neighbor graceful restart active"); -- 2.17.1

2 1

bird remote control - socket unexpectedly closed
by Kenth Eriksson 29 Apr '19

29 Apr '19

Hi! We are remote controlling a bird daemon using a unix domain socket. We have stumbled into an issue where it looks as bird closes the socket (application gets POLLHUP). Any known issues around this area? The pplication and bird daemon continues to run after this happens, but the socket is down. Thanks, Kenth

2 3

Bird 1.6.3 path prepending ?
by Pieter De Wit 27 Apr '19

27 Apr '19

Hi Everyone, I have been fighting with this now for a while and I can’t seem to get my paths prepended. I am running 1.6.3 from the Ubuntu repo. Network wise, I have a VPN server that takes VPNs from remote sites. This VPN server exports the ranges to the rest of the network. It learns the connected sites from a Raspberry PI on site. These links are used as backup paths :) Here is the BGP config on the server, from a remote site: The filters basically just restrict it to RFC1918 addresses protocol bgp bgp_site_1 { description “site_1"; local as 65001; neighbor 10.255.255.101 as 65101; import filter remote_sites; export filter export_to_remote_sites; } Here is the config on the remote site: filter prepend_path { bgp_path.prepend(65101); bgp_path.prepend(65101); bgp_path.prepend(65101); accept; } protocol bgp itc_vpn { description "ITC VPN"; local as 65101; neighbor 10.255.255.1 as 65001; import all; export filter prepend_path; } What am I missing ? Thanks !

2 4

BIRD 2.0.4 segfaulting on ARM
by lorenz＠irmhil.de 26 Apr '19

26 Apr '19

Hello bird-users, bird crashes a few seconds after startup on my Hardkernel ODROID HC-2 running Ubuntu. Kernel version is 4.14.111-158. How can I debug this? I tried: $ bird -c bird.conf -d -D debug.log --> debug.log is 0 bytes. next try: $ bird -c bird.conf $ birdc # debug all all # echo all Last lines of output is: >>> MyOSPF: HELLO packet sent via eth0.1000 >>> MyOSPF: HELLO packet received from nbr 10.99.0.1 on eth0.1000 >>> MyOSPF: HELLO packet received from nbr 10.2.0.0 on eth0.1000 >>> MyOSPF: LSACK packet sent via eth0.1000 >>> MyOSPF: length 56 >>> MyOSPF: router 33.3.0.0 >>> MyOSPF: LSA Type: 4005, Id: 0.0.0.1, Rt: 10.99.0.1, Seq: 8000068a, Age: 1, Sum: 97a9 >>> MyOSPF: LSA Type: 0008, Id: 0.0.0.40, Rt: 10.99.0.1, Seq: 80000633, Age: 1, Sum: e252 Connection closed by server I would be glad helping fixing this bug. Thanks a lot, Lorenz

4 10

Re: BIRD 2.0.4 segfaulting on ARM
by Danny Wensley 25 Apr '19

25 Apr '19

Hi, Just done a bit of research - this closely relates to the main platform I use bird on - and the ODROID HC2 uses a Samsung Exynos5422 processer, which is ArmV7-A. Using qemu-system-arm with the raspi2 machine and a recent version of Raspbian would yield an emulated system with near identical architectural properties as the ODROID HC2. Danny. On Thu, Apr 25, 2019, 11:55 PM Maria Matějka <jan.matejka(a)nic.cz> wrote: > Or, probably what would be even better – could you please help me running > QEMU with this specific guest architecture? This is definitely an > architecture which we would like to include in our test suite. > > Thanks > Maria > > On April 25, 2019 11:36:11 PM GMT+02:00, "Maria Matějka" < > jan.matejka(a)nic.cz> wrote: >> >> Hello! >> >> ulimit -c unlimited >> bird ... >> (after it crashes) >> gdb bird core <<EOF >> backtrace full >> EOF >> >> Or, alternatively, try running it in Valgrind. >> Maria >> >> On April 25, 2019 7:50:32 PM GMT+02:00, lorenz(a)irmhil.de wrote: >>> >>> Hello bird-users, >>> >>> bird crashes a few seconds after startup on my Hardkernel ODROID HC-2 >>> running Ubuntu. Kernel version is 4.14.111-158. >>> >>> >>> How can I debug this? >>> >>> >>> I tried: >>> >>> >>> $ bird -c bird.conf -d -D debug.log >>> >>> --> debug.log is 0 bytes. >>> >>> >>> >>> next try: >>> >>> $ bird -c bird.conf >>> >>> $ birdc >>> >>> # debug all all >>> >>> # echo all >>> >>> Last lines of output is: >>> >>> MyOSPF: HELLO packet sent via eth0.1000 >>>>>> MyOSPF: HELLO packet received from nbr 10.99.0.1 on eth0.1000 >>>>>> MyOSPF: HELLO packet received from nbr 10.2.0.0 on eth0.1000 >>>>>> MyOSPF: LSACK packet sent via eth0.1000 >>>>>> MyOSPF: length 56 >>>>>> MyOSPF: router 33.3.0.0 >>>>>> MyOSPF: LSA Type: 4005, Id: 0.0.0.1, Rt: 10.99.0.1, Seq: >>>>>> >>>>> 8000068a, Age: 1, Sum: 97a9 >>> >>>> MyOSPF: LSA Type: 0008, Id: 0.0.0.40, Rt: 10.99.0.1, Seq: >>>>>> >>>>> 80000633, Age: 1, Sum: e252 >>> Connection closed by server >>> >>> >>> I would be glad helping fixing this bug. >>> >>> >>> Thanks a lot, >>> >>> Lorenz >>> >>> > -- > Sent from my Android device with K-9 Mail. Please excuse my brevity. >

1 0

BIRDv2 OSPF: Stub for loopback potentially broken: Invalid Prefix in LSA
by bauen1 25 Apr '19

25 Apr '19

Hello, I've been playing around with BIRD and OSPF for a while now and ran across an error I don't quite understand, while trying to get BIRD to advertise ip addresses on the loopback interface on a debian 10 server. # bird.conf: debug protocoll all; protocol device { scan time 10; }; protocol ospf v3 ospf_ip4 { ipv4 { export filter accept; import filter accept; import keep filtered; }; area 0.0.0.0 { interface "lo", "dummy*" { stub yes; }; }; }; # ip addr show dev lo 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet 10.50.0.1/32 scope global lo valid_lft forever preferred_lft forever inet6 fe80::1/128 scope link valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever Please note that loopback has a ipv6 link local address to get bird to recognize the interface. In theory, this should advertise 10.50.0.1/32 to other ospf routers, yet it doesn't and BIRD complains about LSAs with an invalid prefix This should advertise 10.50.0.1/32 to other routers and add the respective entries to the LSADB, yet bird starts complaing about an invalid LSA: "ospf_ip4: Invalid prefix in LSA (Type: 2001, Id: 0.0.0.0, Rt: 10.50.0.1)", I have attached the debug output of bird. Any help would be appreciated.

4 15