Hello,

Just upgrading to BB. There is something wrong with multicast/igmpproxy during dhcp renew on wan:

Wed Nov 12 11:15:22 2014 daemon.notice netifd: wan (1291): Sending renew...
Wed Nov 12 11:15:22 2014 daemon.notice netifd: wan (1291): Lease of 77.207.xx.xx obtained, lease time 180
Wed Nov 12 11:15:22 2014 user.warn igmpproxy[23620]: select() failure; Errno(4): Interrupted system call

This behavior causes an igmp leave on lan:

11:15:22 .567273 IP openwrt > all-routers.mcast.net: igmp leave all-routers.mcast.net
11:15:22 .905350 IP openwrt > all-systems.mcast.net: igmp query v2

As result, xcast TV is down for few seconds.

I've read somewhere that netifd has been changed in BB, maybe it is related... Everything was fine on AA.

I need some help, as you can see, my isp gives me very short lease: I can't watch TV anymore.

(Last edited by nague on 12 Nov 2014, 11:34)

Actually, igmpproxy is not crashing but being restart for every dhcp renew on wan.

I think this behavior is related to this changset: https://dev.openwrt.org/changeset/41011

In fact, igmpproxy should only be reloaded for ifup/ifdown events, not ifupdate (dhcp). I don't know where to check that (I'm not familiar with the new netifd) but removing theses lines from /etc/init.d/igmpproxy solved my issue:

#       procd_open_trigger
#       config_foreach igmp_add_network phyint
#       procd_close_trigger

(Last edited by nague on 12 Nov 2014, 16:49)

Observed on BB here too. Made a ticket for that. https://dev.openwrt.org/ticket/20028

This has reappeared in Chaos Calmer:

Mon Feb 29 13:17:08 2016 daemon.notice netifd: wan (1254): Lease of XX.XX.XX.XXX obtained, lease time 900
Mon Feb 29 13:17:08 2016 user.warn igmpproxy[18573]: select() failure; Errno(4): Interrupted system call