We are on a big project, making IPTV service for our clients based on Mikrotik RBs. Total project includes about 80 streams, 250 RB800 (one on every building), 2000+ clients Scenario: 1 IPTV Streamer 1 RouterBoard800 2 Clients - 2 STB (ether 10, 12 and 13)
Everything works like a charm except 1 big issue. Problem: Client A(first STB) joins 1st channel / 1 stream goes to interface eth10. Client B(second STB) joins 1st channel / same stream goes to interface eth 12.
So far so good.
Clients B(ether12) leaves 1st channel and goes to 2nd channel / 2 streams goes to eth12 - trick is that 1st channel stream doesnt stop till client A (STB/ether 10) doesnt leave 1st channel. Exception is that Client B leaves every stream except streams already being recieved on any other interface/STBs.
Theoreticaly if we got 40 Clients/Interfaces and 40 streams/mcast groups, 1 of clients could end up with 40 streams hitting his interface and he is only watching 1.
Somehow client's(on 1 interface) 'leave message', who is leaving channel already watched by other clients (on other interface), is beeing ignored on same router for that client's interface.
We tried this scenario with few versions of Ros and devices. PIM/IGMP isn't option for us.
Log:
Code:
01:50:16 igmp-proxy,debug RECV IGMP leave message from 1.10.0.254 to 224.0.0.2 on ether2-stb 01:50:16 igmp-proxy,debug leaving multicast group 232.100.3.2 on ether1-ulaz 01:50:16 igmp-proxy,debug sending IGMP query to 232.100.3.2 on ether2-stb 01:50:16 igmp-proxy,debug RECV IGMP membership query from 178.209.0.85 to 232.100. 3.2 on ether1-ulaz 01:50:16 igmp-proxy,debug ignoring IGMP message: received on the upstream interfac e 01:50:16 igmp-proxy,debug RECV IGMPv2 membership report from 1.10.0.254 to 232.100 .6.2 on ether2-stb 01:50:16 igmp-proxy,debug joining multicast group 232.100.6.2 on ether1-ulaz 01:50:16 igmp-proxy,debug ignoring IGMP message: source address is local: 01:50:16 igmp-proxy,debug source=141.170.196.139 01:50:16 igmp-proxy,debug interface=ether1-ulaz 01:50:16 igmp-proxy,debug ignoring request from myself: 01:50:16 igmp-proxy,debug source=141.170.196.139 01:50:16 igmp-proxy,debug destination=232.100.6.2 01:50:16 igmp-proxy,debug received notification: 01:50:16 igmp-proxy,debug source=141.170.196.129 01:50:16 igmp-proxy,debug destination=232.100.6.2 01:50:16 igmp-proxy,debug adding multicast forwarding entry 01:50:16 igmp-proxy,debug group: 232.100.6.2 01:50:16 igmp-proxy,debug source: 141.170.196.129 01:50:17 igmp-proxy,debug RECV IGMP membership query from 178.209.0.85 to 232.100. 3.2 on ether1-ulaz 01:50:17 igmp-proxy,debug ignoring IGMP message: received on the upstream interfac e 01:50:18 igmp-proxy,debug RECV IGMPv2 membership report from 1.10.0.254 to 232.100 .6.2 on ether2-stb 01:50:18 igmp-proxy,debug sending IGMP query to 232.100.3.2 on ether2-stb 01:50:20 igmp-proxy,debug RECV IGMP leave message from 1.10.0.254 to 224.0.0.2 on ether2-stb 01:50:20 igmp-proxy,debug leaving multicast group 232.100.6.2 on ether1-ulaz 01:50:20 igmp-proxy,debug sending IGMP query to 232.100.6.2 on ether2-stb 01:50:20 igmp-proxy,debug RECV IGMP membership query from 178.209.0.85 to 232.100. 6.2 on ether1-ulaz 01:50:20 igmp-proxy,debug ignoring IGMP message: received on the upstream interfac e 01:50:20 igmp-proxy,debug RECV IGMPv2 membership report from 1.10.0.254 to 232.100 .6.3 on ether2-stb 01:50:20 igmp-proxy,debug adding multicast forwarding entry 01:50:20 igmp-proxy,debug group: 232.100.6.3 01:50:20 igmp-proxy,debug source: 141.170.196.129 01:50:20 igmp-proxy,debug adding multicast forwarding entry 01:50:20 igmp-proxy,debug group: 232.100.6.3 01:50:20 igmp-proxy,debug source: 178.209.0.85 01:50:21 igmp-proxy,debug RECV IGMPv2 membership report from 1.10.0.254 to 232.100 .6.3 on ether2-stb 01:50:21 igmp-proxy,debug RECV IGMP membership query from 178.209.0.85 to 232.100. 6.2 on ether1-ulaz 01:50:21 igmp-proxy,debug ignoring IGMP message: received on the upstream interfac e
Hey Ropeba ? Would you mind doing the following: /routing igmpproxy print and past the result there ? Thank you
ropeba wrote:
We are on a big project, making IPTV service for our clients based on Mikrotik RBs. Total project includes about 80 streams, 250 RB800 (one on every building), 2000+ clients Scenario: 1 IPTV Streamer 1 RouterBoard800 2 Clients - 2 STB (ether 10, 12 and 13)
Everything works like a charm except 1 big issue. Problem: Client A(first STB) joins 1st channel / 1 stream goes to interface eth10. Client B(second STB) joins 1st channel / same stream goes to interface eth 12.
So far so good.
Clients B(ether12) leaves 1st channel and goes to 2nd channel / 2 streams goes to eth12 - trick is that 1st channel stream doesnt stop till client A (STB/ether 10) doesnt leave 1st channel. Exception is that Client B leaves every stream except streams already being recieved on any other interface/STBs.
Theoreticaly if we got 40 Clients/Interfaces and 40 streams/mcast groups, 1 of clients could end up with 40 streams hitting his interface and he is only watching 1.
Somehow client's(on 1 interface) 'leave message', who is leaving channel already watched by other clients (on other interface), is beeing ignored on same router for that client's interface.
We tried this scenario with few versions of Ros and devices. PIM/IGMP isn't option for us.
Log:
Code:
01:50:16 igmp-proxy,debug RECV IGMP leave message from 1.10.0.254 to 224.0.0.2 on ether2-stb 01:50:16 igmp-proxy,debug leaving multicast group 232.100.3.2 on ether1-ulaz 01:50:16 igmp-proxy,debug sending IGMP query to 232.100.3.2 on ether2-stb 01:50:16 igmp-proxy,debug RECV IGMP membership query from 178.209.0.85 to 232.100. 3.2 on ether1-ulaz 01:50:16 igmp-proxy,debug ignoring IGMP message: received on the upstream interfac e 01:50:16 igmp-proxy,debug RECV IGMPv2 membership report from 1.10.0.254 to 232.100 .6.2 on ether2-stb 01:50:16 igmp-proxy,debug joining multicast group 232.100.6.2 on ether1-ulaz 01:50:16 igmp-proxy,debug ignoring IGMP message: source address is local: 01:50:16 igmp-proxy,debug source=141.170.196.139 01:50:16 igmp-proxy,debug interface=ether1-ulaz 01:50:16 igmp-proxy,debug ignoring request from myself: 01:50:16 igmp-proxy,debug source=141.170.196.139 01:50:16 igmp-proxy,debug destination=232.100.6.2 01:50:16 igmp-proxy,debug received notification: 01:50:16 igmp-proxy,debug source=141.170.196.129 01:50:16 igmp-proxy,debug destination=232.100.6.2 01:50:16 igmp-proxy,debug adding multicast forwarding entry 01:50:16 igmp-proxy,debug group: 232.100.6.2 01:50:16 igmp-proxy,debug source: 141.170.196.129 01:50:17 igmp-proxy,debug RECV IGMP membership query from 178.209.0.85 to 232.100. 3.2 on ether1-ulaz 01:50:17 igmp-proxy,debug ignoring IGMP message: received on the upstream interfac e 01:50:18 igmp-proxy,debug RECV IGMPv2 membership report from 1.10.0.254 to 232.100 .6.2 on ether2-stb 01:50:18 igmp-proxy,debug sending IGMP query to 232.100.3.2 on ether2-stb 01:50:20 igmp-proxy,debug RECV IGMP leave message from 1.10.0.254 to 224.0.0.2 on ether2-stb 01:50:20 igmp-proxy,debug leaving multicast group 232.100.6.2 on ether1-ulaz 01:50:20 igmp-proxy,debug sending IGMP query to 232.100.6.2 on ether2-stb 01:50:20 igmp-proxy,debug RECV IGMP membership query from 178.209.0.85 to 232.100. 6.2 on ether1-ulaz 01:50:20 igmp-proxy,debug ignoring IGMP message: received on the upstream interfac e 01:50:20 igmp-proxy,debug RECV IGMPv2 membership report from 1.10.0.254 to 232.100 .6.3 on ether2-stb 01:50:20 igmp-proxy,debug adding multicast forwarding entry 01:50:20 igmp-proxy,debug group: 232.100.6.3 01:50:20 igmp-proxy,debug source: 141.170.196.129 01:50:20 igmp-proxy,debug adding multicast forwarding entry 01:50:20 igmp-proxy,debug group: 232.100.6.3 01:50:20 igmp-proxy,debug source: 178.209.0.85 01:50:21 igmp-proxy,debug RECV IGMPv2 membership report from 1.10.0.254 to 232.100 .6.3 on ether2-stb 01:50:21 igmp-proxy,debug RECV IGMP membership query from 178.209.0.85 to 232.100. 6.2 on ether1-ulaz 01:50:21 igmp-proxy,debug ignoring IGMP message: received on the upstream interfac e
Hi, right now we have deployed over 250 GEPON ONUs which don't support PIM. Our only solution is to use IGMP proxy. (we dont have this problem with PIM, but we do have othr problem with GEPON and PIM, so the only thing left is IGMP proxy)
Joined: Tue Feb 14, 2006 9:46 am Posts: 4811
Karma: 37
Location: Riga, Latvia
PIM explicitly sets what IGMP version to use, by default, it is IGMPv2, you can try to set other IGMP version there, most probably IGMPv1
IGMP-proxy uses auto-configuration to determine what version of IGMP to use. that is why you could have problem with PIM and not with IGMP-Proxy.
Also, are you sure that your devices are sending leave (larger section or relevant Log messages could be helpful) just open new temrinal and there issue "/log print follow" check if RouterOS is receiving leave messages correctly.
Also, if source of stream is in same network as interface address multicast is running on, you can skip setting up alternative-networks, also, try adding only source address network there.
let's begin from start. We must to use IGMP Proxy, as i said before, PIM is not an option (GEPON limitation) . This is simple configuration whitch doesn't work OK with mikrotik IGMP proxy.
In this case on STB1 I play "channe 1", on STB2 play "channel 1" and everything is ok. After that i change chanel on STB1 to "chanel 2" and I watch that channel without problem. But, traffic of "channel 1" still hitting interface of STB1 and it will not stop till I change channel on STB2 to any other. It's simple configuration of IGMP proxy, and i doubt that I can make any mistake. That looks like a bug in IGMP proxy in ROS.
log is here
Code:
23:16:27 igmp-proxy,debug RECV IGMP leave message from 1.220.0.2 to 224.0.0.2 on ether10-stb1
23:16:27 igmp-proxy,debug sending IGMP query to 239.100.6.5 on ether10-stb1
23:16:27 igmp-proxy,debug RECV IGMPv2 membership report from 1.220.0.2 to 239.100.6.8 on ether10-stb1
23:16:27 igmp-proxy,debug joining multicast group 239.100.6.8 on ether2-ulaz
23:16:27 igmp-proxy,debug ignoring IGMP message: source address is local:
23:16:27 igmp-proxy,debug source=192.168.84.230
23:16:27 igmp-proxy,debug interface=ether2-ulaz
23:16:27 igmp-proxy,debug ignoring request from myself:
We have built a complete FTTb network in a city with 154 buildings. We put RB493G in every entrance of the building and every user has it's own port on that RB. Internet service is working without any problems for some time, but our only solution for IPTV service is IGMP proxy. Replacing GPON terminal equipment is not an option.
Joined: Sun Dec 04, 2011 11:32 pm Posts: 5
Karma: 0
Location: Odessa, Ukraine
Hi all! I have the same problem as people above. RouterOS x86 v5.9 IGMP-Proxy says in log: "ignoring request from myself" and in MFC table "Unknown" interface as "source" appears. like:
Joined: Sat Feb 27, 2010 6:13 pm Posts: 35
Karma: 1
I had similar problems and error messages until I ran
/routing igmp-proxy interface print status
and noticed that the source-ip-address for the upstream and downstream interfaces were the same. This was due to me having a 1.2.3.4/29 address on ether2 and also 1.2.3.4/32 on pptp-out1 (pseudo unnumbered interface). IGMP proxy looks like it takes the IP address configured on the relevant interfaces and uses them. Since both were the same it appears to get confused.
on igmp proxy server you need to add all streamer ip subnet to alternative subnet on igmp proxy settings. All works ok, many my clients use it.
Read the complete topic in detail, try the same scenario as in me and you'll see that you are wrong. And that has nothing to do with alternative subnets. I have FTTB network in which I'm using mikrotik RB as access equipment in buildings. If in a building 8 users is connected on 8 ports at the RB 493g, and if 7 users is watching 7 different channels, whenever the eighth user starts some of those channels and switch on some other channel, those 7 will continue to go to his STB and even if he want to watch only one channel the stream of unwanted channels will not stop until everyone in the building go out of the stream channel. I have a scenario where a user goes 50 Mbit of traffic, and he's watching only one Channel of 5 mbit.s I have FTTB network with nearly 300 mikrotik Routerboards and at each locations behaves identically, unnecessarily overloading RB, STB and all other equipment on the user side.
Joined: Tue Feb 14, 2006 9:46 am Posts: 4811
Karma: 37
Location: Riga, Latvia
waiting for new routing to come to RouterOS. That should fix a lot of underlying issues. After that problems with pim and igmp are going to be dealt with
waiting for new routing to come to RouterOS. That should fix a lot of underlying issues. After that problems with pim and igmp are going to be dealt with
looking forward to the MUM... *crossing fingers*
_________________ For every complex problem, there is a solution that is simple, neat, and wrong.
Joined: Sun Dec 04, 2011 11:32 pm Posts: 5
Karma: 0
Location: Odessa, Ukraine
Yes! IGMP Proxy works for me.
After hours of magic i've set igmp proxying right. But streams stop after some minutes of viewing (seems to be a response waiting time). In Wireshark sniffer i saw address 254.0.16.172 (while my "LAN" interface address was 172.16.0.254) sending group membership requests but my PC didn't response on it. Seems like igmp proxy mirrors ip address. Another try was to set IP address to something else. And Oh! YEAH! setting address of LAN interface to 172.16.0.1 was good. Igmp proxy still mirrors IP address to 1.0.16.172 but my PC send group packets back and playing don't stop.
Is this any kind of feature?
_________________ Resurrecting the dead
Last edited by explorer84 on Thu Apr 05, 2012 4:42 pm, edited 1 time in total.
In addition to the great problems which I described in this topic, IGMP proxy works ok, we got test it to the limit of we have found no other flaws. We have a full functional FTTB network which offer Internet access services and IP television with over 80 TV channels. On this occasion I wanted to ask how far is the solutions to this problem, almost a year to be those bug is reported. As for our network, It's already a critical situation that each RouterBOARD in the entrance of a building is full (users connected to almost all on every ethernet ports) and the CPUs on the RB are working non-stop at 100% and there is a problem in the IPTV service, braking, etc, and the reason is those useless 50-100 Mbit on every ethernet interface where it realistically should be 5-6 Mbps. If mikrotik soon not solve this problem, we will have to migrate a complete access equipment in buildings to another platform, on which IGMP proxy works ok. Yet, I hope that we can avoid that and that Mikrotik will fix that up.
Joined: Sun Dec 04, 2011 11:32 pm Posts: 5
Karma: 0
Location: Odessa, Ukraine
Chupaka wrote:
explorer84 wrote:
Seems like igmp proxy mirrors ip address.
what version do you use?
I use version 5.14, but on 4.17 it does the same (tested on VMWare)
also I figured out that setting "querier" IP address to x.x.x.224 and higher in the same subnet (x.x.x.225 and so on) also makes PC not to answer on query packets. PC answers on "query" packets if "querier" IP address set to something beetween x.x.x.1 and x.x.x.223. (Windows 7 and Windows XP work the same).
Users browsing this forum: No registered users and 2 guests
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot post attachments in this forum
Login |
HOME
