NTP listen broadcast IP

Question

We have restricted NTP to listen only eth0 and rest should will ignore. However we can see that after making changes in ntp.conf it is trying to listen broadcast and failing to bind address with unexpected error.

ntpd[89217]: ./../lib/isc/unix/ifiter_ioctl.c:617: unexpected error:
ntpd[89217]: eth2:6: getting broadcast address: Cannot assign requested address
ntpd[89217]: i/o error on routing socket No buffer space available – disabling
ntpd[5410]: ./../lib/isc/unix/ifiter_ioctl.c:617: unexpected error:
ntpd[5410]: eth3.1238:0: getting broadcast address: Cannot assign requested address
ntpd[5410]: ntpd exiting on signal 15
ntpd[1508]: ntpd exiting on signal 15

Below is the ntp.conf from customer's server :-

cat ntp.conf

fudge 127.127.1.0 stratum 10

Authentication stuff

keys /etc/ntp.keys
path for keys file
trustedkey 1
define trusted keys
requestkey 1
server 172.23.5.8 iburst
server 172.23.5.9 iburst
restrict 172.23.5.8
restrict 172.23.5.9
key (7) for accessing server variables
controlkey 15 # key (6) for accessing server variables
extra lines to fix issue about NTP Daemon
interface listen eth0
interface ignore wildcard

This one is /var/lib/NTP log file you can see.

27 Sep 10:30:17 ntpd[89217]: ./../lib/isc/unix/ifiter_ioctl.c:617: unexpected error:
27 Sep 10:30:17 ntpd[89217]: eth2:6: getting broadcast address: Cannot assign requested address
27 Sep 10:30:17 ntpd[89217]: i/o error on routing socket No buffer space available - disabling
11 Oct 13:12:46 ntpd[5410]: ./../lib/isc/unix/ifiter_ioctl.c:617: unexpected error:
11 Oct 13:12:46 ntpd[5410]: eth3.1238:0: getting broadcast address: Cannot assign requested address
11 Oct 13:13:23 ntpd[5410]: ./../lib/isc/unix/ifiter_ioctl.c:617: unexpected error:
11 Oct 13:13:23 ntpd[5410]: eth3.1277:0: getting broadcast address: Cannot assign requested address
11 Oct 13:13:41 ntpd[5410]: i/o error on routing socket No buffer space available - disabling
13 Oct 21:11:05 ntpd[5410]: ntpd exiting on signal 15
13 Oct 20:50:18 ntpd[1508]: ntpd exiting on signal 15
13 Oct 22:49:52 ntpd[113345]: LOCAL(0) 8011 81 mobilize assoc 35168
13 Oct 22:49:52 ntpd[113345]: 172.23.5.8 8011 81 mobilize assoc 35169
13 Oct 22:49:52 ntpd[113345]: 172.23.5.9 8011 81 mobilize assoc 35170
13 Oct 22:49:52 ntpd[113345]: 0.0.0.0 c016 06 restart
13 Oct 22:49:52 ntpd[113345]: 0.0.0.0 c012 02 freq_set kernel -22.705 PPM
13 Oct 20:49:53 ntpd[113345]: LOCAL(0) 8024 84 reachable
13 Oct 20:49:53 ntpd[113345]: LOCAL(0) 963a 8a sys_peer
13 Oct 20:49:53 ntpd[113345]: 0.0.0.0 c515 05 clock_sync
13 Oct 20:49:54 ntpd[113345]: 172.23.5.8 8024 84 reachable
13 Oct 20:49:55 ntpd[113345]: 172.23.5.9 8024 84 reachable
13 Oct 20:50:57 ntpd[113345]: 172.23.5.9 963a 8a sys_peer
13 Oct 20:53:18 ntpd[113345]: LOCAL(0) 961a 8a sys_peer
13 Oct 20:59:48 ntpd[113345]: ntpd exiting on signal 15
13 Oct 22:59:40 ntpd[122930]: LOCAL(0) 8011 81 mobilize assoc 21051
13 Oct 22:59:40 ntpd[122930]: 172.23.5.8 8011 81 mobilize assoc 21052
13 Oct 22:59:40 ntpd[122930]: 172.23.5.9 8011 81 mobilize assoc 21053
13 Oct 22:59:40 ntpd[122930]: 0.0.0.0 c016 06 restart
13 Oct 22:59:40 ntpd[122930]: 0.0.0.0 c012 02 freq_set kernel -22.705 PPM
13 Oct 20:59:41 ntpd[122930]: LOCAL(0) 8024 84 reachable
13 Oct 20:59:41 ntpd[122930]: LOCAL(0) 963a 8a sys_peer
13 Oct 20:59:41 ntpd[122930]: 0.0.0.0 c515 05 clock_sync
13 Oct 21:00:39 ntpd[122930]: ntpd exiting on signal 15
13 Oct 23:00:39 ntpd[124181]: LOCAL(0) 8011 81 mobilize assoc 30350
13 Oct 23:00:39 ntpd[124181]: 172.23.5.8 8011 81 mobilize assoc 30351
13 Oct 23:00:39 ntpd[124181]: 172.23.5.9 8011 81 mobilize assoc 30352
13 Oct 23:00:39 ntpd[124181]: 0.0.0.0 c016 06 restart
13 Oct 23:00:39 ntpd[124181]: 0.0.0.0 c012 02 freq_set kernel -22.705 PPM
13 Oct 21:00:40 ntpd[124181]: LOCAL(0) 8024 84 reachable
13 Oct 21:00:40 ntpd[124181]: LOCAL(0) 963a 8a sys_peer
13 Oct 21:00:40 ntpd[124181]: 0.0.0.0 c515 05 clock_sync
13 Oct 21:00:41 ntpd[124181]: 172.23.5.8 8024 84 reachable
13 Oct 21:00:42 ntpd[124181]: 172.23.5.9 8024 84 reachable
13 Oct 21:01:44 ntpd[124181]: 172.23.5.9 963a 8a sys_peer
13 Oct 21:05:11 ntpd[124181]: LOCAL(0) 961a 8a sys_peer
13 Oct 21:16:23 ntpd[124181]: i/o error on routing socket No buffer space available - disabling
13 Oct 21:28:42 ntpd[124181]: ntpd exiting on signal 15
13 Oct 23:32:35 ntpd[5438]: LOCAL(0) 8011 81 mobilize assoc 54508
13 Oct 23:32:35 ntpd[5438]: 172.23.5.8 8011 81 mobilize assoc 54509
13 Oct 23:32:35 ntpd[5438]: 172.23.5.9 8011 81 mobilize assoc 54510
13 Oct 23:32:35 ntpd[5438]: 0.0.0.0 c016 06 restart
13 Oct 23:32:35 ntpd[5438]: 0.0.0.0 c012 02 freq_set kernel -22.705 PPM
13 Oct 21:32:36 ntpd[5438]: LOCAL(0) 8024 84 reachable
13 Oct 21:32:36 ntpd[5438]: LOCAL(0) 963a 8a sys_peer
13 Oct 21:32:36 ntpd[5438]: 0.0.0.0 c515 05 clock_sync
13 Oct 21:32:37 ntpd[5438]: 172.23.5.8 8024 84 reachable
13 Oct 21:32:38 ntpd[5438]: 172.23.5.9 8024 84 reachable
13 Oct 21:33:40 ntpd[5438]: 172.23.5.9 963a 8a sys_peer
13 Oct 21:37:05 ntpd[5438]: 172.23.5.8 943d 8d popcorn 0.011376 s
13 Oct 21:40:30 ntpd[5438]: 172.23.5.9 964d 8d popcorn 0.013156 s
13 Oct 21:42:12 ntpd[5438]: LOCAL(0) 8043 83 unreachable

Hope you got something.

Can any body tell me how to fix this error?

We'd need to see the whole of your `ntp.conf`, and, if at all possible, the unredacted output of `ip a sh`. — MadHatter, Oct 20 '16 at 08:03
@MadHatter my ip a sh output is very lengthy. is there any way to send you. This is SBC machine, and eth0 is a real interface. but i do not know from where NTP had got eth2:6 and eth3:1238:0. I have defined this interface too but did not include in ntp.conf. eth2 and eth3 are using as a signaling interface. — Sumit Gemini, Oct 21 '16 at 10:25
Fair comment. Are you saying that this machine doesn't have the interfaces it's complaining about (eth2:6, eth3.1238:0)? — MadHatter, Oct 21 '16 at 10:38
@MadHatter i am working on SBC machine (session border controller). Here we have created eth2 and eth3 as a signaling interface. eth2 and eth3 are treated as a vlan and bind relem eth2:6 and eth3.1238:0 to these vlan. In are case we discard eth2 and eth3 from ntp.conf because there are multiple relem bind to eth2 and eth3 so ntp try to create socket for each session. and problem was File descriptor exhaust. that's why we add only eth0 interface. Hope you understand. — Sumit Gemini, Oct 21 '16 at 10:49
@MadHatter could please tell me, in which scenario, i get following error :- 1) i/o error on routing socket No buffer space available – disabling 2) ntpd exiting on signal 15 — Sumit Gemini, Oct 22 '16 at 07:32
My hypothesis was that your sheer number of interfaces was overrunning some internal buffer in the daemon. But I've gone away and tested the use of `interface listen eth0`, and it seems to work for me; it may be that there's still an overrun when ntpd itemises all the interfaces, even though it doesn't want to bind to them. Any chance of restarting ntpd and posting the logs? — MadHatter, Oct 22 '16 at 07:58
@MadHatter here i copied ntp log. and did not observe shutdown and start kind of logs in /var/log/message, so can not say ntpd was rebooted or not. if you share your mail-id i can share all logs. — Sumit Gemini, Oct 22 '16 at 08:09
You need to find the rest of the logs. If that means you have to learn about rsyslog, and reconfigure yours, so be it - but the logs are where the software gets to tell you what the problem might be, and you need them. In passing, could I ask you what's in `/proc/swaps`? — MadHatter, Oct 22 '16 at 08:22
Let us [continue this discussion in chat](http://chat.stackexchange.com/rooms/47213/discussion-between-sumit-gemini-and-madhatter). — Sumit Gemini, Oct 22 '16 at 08:25

NTP listen broadcast IP

0 Answers0