Working Link Aggregation Config RT-AC66U

[KAD]

myself I dont have a synology nas

but if you look a page or two back, somebody listed a specific DSM version that was working

 

[ebalsumgo]

right. I am running latest as well and it seems no matter what I do when i run the bonding script all I get is that anything that is hooked up to the wired ports dies until i reboot the router and still no bonding happens :/

 

[KAD]

what's does LinkAgg status option return

LinkAgg -s
or
LinkAgg --status

also you're on AC68U
I would assume RMerlin included the bonding driver in the build for the new model, but you can check that module is loaded with

lsmod | grep -i bond

 

[ebalsumgo]

It doesnt return. When it runs it runs for a sec and then drops two messages that it deleted dead vlans and thats it. I checked and the bonding driver is there and loading.

 

[KAD]

it doesn't sound like your running the status command

dead vlans are deleted when you re-run the script and it has to first delete the old bond
then delete the dead vlans, and recreate new vlans and bond them together

status would give you an output of what vlans are used, are they up
are they correctly listed as slaves
is bond0 up
etc ...

note. I'm still at work, I can post example output later tonight

 

[ebalsumgo]

Sorry, I myself am posting from work and did not provide the full info of what's going on. When I type in ./LinkPorts 3 4 I end up with the following:
1. Instant network drop of all wired connections
2. Script returns after 10 secs (as designed) and tells me it has deleted two dead bonds.
3. There are no bonds and now I have to reboot the router to get the wired connections back up.

I am running the latest dsm, latest Merlin(stable). I have tried with static ip on synology, jumbo frames disabled on both devices, as well as dhcp and jumbo frames enabled. I do have spanning tree enabled.

 

[KAD]

ok, so we'll have to do a little digging to find out what is failing

so clarify my above post there are actually 2 possiblities for deleting a dead vlan

1. on first run attempt
script creates vlans
creates bond
and finally runs a check to make sure everything was successful
if it was not successfull
it will delete the dead - non-function vlans

2. senario 2
user already has an active bond
and re-runs the script
script will recognize that a bond already exist
so it will delete the existing bond
delete the dead - non-functioning vlans
then set up new vlans and bond with new args given
and finally check the status again
and yes if check fails, it will again delete dead vlans

probably the easiest way to find out where it is failing

sh -xv /LinkAgg 3 4 >> /path/LinkAgg.log 2>&1

this will run the script with a massive amount of logging for each
you will not see the output on terminal, it will redirect all the output to a log file, you can specify the location
I'd be happy to review you can post the contents up here, or post the file somewhere that I can review it

 

[ebalsumgo]

It never creates a bond, so I am pretty sure it is option 1. I will run it in debug mode and post it later tonight.

 

[ebalsumgo]

Ok here is what we got: http://pastebin.com/haJ0MHXs

 

[KAD]

thanks for the file

a very quick look, it says that the ports are not up and running
I'll be taking a closer look a bit later tonight

something interesting that might be a quick fix
after creating a bond it issues
sleep 10
I selected 10 because it takes a few seconds for everything to come up
in all of my test everything was up in approx 5 - 6 seconds
I figured 10 would cover the basis,
but I wonder if you tried 15 or 20, if that would make any difference

anyways, some family time now, will look closer later

and as mentioned earlier, here's some output example of status cmd

xx@RT-AC66U:/tmp/mnt/sda1# ./LinkAgg -s

--- Bond Errors ---





--- Bond Status ---

Bond Status: bond0 UP
Bridge to WAN Status: Member of br0=bond0
Slave 1 Status: vlan=vlan3 Link=UP Port=3
Slave 2 Status: vlan=vlan4 Link=UP Port=4

xx@RT-AC66U:/tmp/mnt/sda1#

there's no errors
and all status items look good

 

[KAD]

@ebalsumgo,

please do the following and post output

robocfg vlan 1 ports "1 2 8t"
robocfg vlan 3 ports "3 8t"
robocfg vlan 4 ports "4 8t"
vconfig add eth0 3
vconfig add eth0 4
echo 802.3ad > /sys/class/net/bond0/bonding/mode
echo fast > /sys/class/net/bond0/bonding/lacp_rate
ip link set bond0 up
echo 100 > /sys/class/net/bond0/bonding/miimon
echo +vlan3 > /sys/class/net/bond0/bonding/slaves
echo +vlan4 > /sys/class/net/bond0/bonding/slaves
brctl addif br0 bond0
iptables -I INPUT 1 -i vlan3 -j ACCEPT
iptables -I INPUT 1 -i vlan4 -j ACCEPT
iptables -I INPUT 1 -i ${1} -j ACCEPT
ip link show

I see an issue with status checker, something I did not expect

this code

Check2=$(ip link show | grep "$CHECKVLAN1" | cut -d '<' -f 2 | grep -o UP)
		if [ "$Check2" != "UP" ]; then

checks vlan status for UP
but in your case it's returning multiple entries

specially
"UP
UP
UP"

which seems odd, there should only be 1 entry per vlan

 

[ebalsumgo]

admin@RT-AC68R:/tmp/home/root# modprobe bonding
admin@RT-AC68R:/tmp/home/root# robocfg vlan 1 ports "1 2 8t"
admin@RT-AC68R:/tmp/home/root# robocfg vlan 3 ports "3 8t"
admin@RT-AC68R:/tmp/home/root# robocfg vlan 4 ports "4 8t"
admin@RT-AC68R:/tmp/home/root# vconfig add eth0 3
admin@RT-AC68R:/tmp/home/root# vconfig add eth0 4
admin@RT-AC68R:/tmp/home/root# echo 802.3ad > /sys/class/net/bond0/bonding/mode
admin@RT-AC68R:/tmp/home/root# echo fast > /sys/class/net/bond0/bonding/lacp_rate
admin@RT-AC68R:/tmp/home/root# ip link set bond0 up
admin@RT-AC68R:/tmp/home/root# echo 100 > /sys/class/net/bond0/bonding/miimon
admin@RT-AC68R:/tmp/home/root# echo +vlan3 > /sys/class/net/bond0/bonding/slaves
admin@RT-AC68R:/tmp/home/root# echo +vlan4 > /sys/class/net/bond0/bonding/slaves
admin@RT-AC68R:/tmp/home/root# brctl addif br0 bond0
admin@RT-AC68R:/tmp/home/root# iptables -I INPUT 1 -i vlan3 -j ACCEPT
admin@RT-AC68R:/tmp/home/root# iptables -I INPUT 1 -i vlan4 -j ACCEPT
admin@RT-AC68R:/tmp/home/root# iptables -I INPUT 1 -i bond0 -j ACCEPT
admin@RT-AC68R:/tmp/home/root# ip link show
1: lo: <LOOPBACK,MULTICAST,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
    link/ether 00:01:36:1f:e4:5a brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
    link/ether 74:d0:2b:65:29:f0 brd ff:ff:ff:ff:ff:ff
4: eth2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UNKNOWN qlen 1000
    link/ether 74:d0:2b:65:29:f4 brd ff:ff:ff:ff:ff:ff
5: vlan1@eth0: <BROADCAST,MULTICAST,PROMISC,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP 
    link/ether 74:d0:2b:65:29:f0 brd ff:ff:ff:ff:ff:ff
6: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN 
    link/ether 74:d0:2b:65:29:f0 brd ff:ff:ff:ff:ff:ff
7: tun21: <NO-CARRIER,POINTOPOINT,MULTICAST,NOARP,PROMISC,UP> mtu 1500 qdisc pfifo_fast state DOWN qlen 100
    link/none 
8: bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP 
    link/ether 00:01:36:1f:e4:5a brd ff:ff:ff:ff:ff:ff
9: vlan3@eth0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc noqueue master bond0 state UP 
    link/ether 00:01:36:1f:e4:5a brd ff:ff:ff:ff:ff:ff
10: vlan4@eth0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc noqueue master bond0 state UP 
    link/ether 00:01:36:1f:e4:5a brd ff:ff:ff:ff:ff:ff
admin@RT-AC68R:/tmp/home/root#

same end result, wired connections all die until i reboot the router

 

[KAD]

well regarding status checker, with the output above I can fix that

see output

9: vlan3@eth0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc noqueue master bond0 state UP

has muliple "UP" maybe due to running vpn
not going to be fixed today, it's turkey day here in the states

but as for the no connection, and needing a reboot
you could try turning off vpn, to see if that makes any difference

 

[vims]

Firstly, Hello and thank you for making this possible with ac66u.

Just installed this together with merlin latest fw 3.0.0.4.374.35_2 and as with Mr. Zoon here I lack the fundamental knowledge to do this with ease.
I've got LAG bonding and it's all and well but for some reason wireless clients can't communicate with NAS. I'm not using any VPN.
My NAS is the synology 412+
Can LAG be causing some issues with wireless?
Wireless clients work otherwise, just not reaching the NAS.

 

[KAD]

unfortunately synology seems to have very limited support for LAG

their official device supported list, includes only 3 switches
that's 3 specific model numbers
and 2 from companies, I had never heard of before

many people are reporting problems with synology NAS using LAG

Here's a quote from a few pages back from TinMan11

I got link aggregation working between my Synology DS412+ (running DSM 4.3-3810) and my Asus RT-N66U (running 3.0.0.4.374.34 (Merlin build))

there's been 2 reports of this working, but only with this very specific synology DSM build

if there's an error is the set up script, such as the one recently found, which will be fixed soon, I can fix items like that

however, what I can not fix is synology's broken implementation of LAG

 

[KAD]

LinkAgg version 1.5 https://www.mediafire.com/?8y3ye3332durchp

@ebalsumgo, please test version 1.5
this should solve the issue with status checker

 

[ebalsumgo]

I will test it out when I get home Sunday from turkey day travels

 

[saintdev]

I wanted to mention I'm having some strange issues with link aggregation. It will work for a little while, then suddenly I will no longer be able to fetch some resources outside my network. For example if I attempt to wget a file off an external HTTP server, it will just hang at awaiting response. Same thing happens with rsync transfers. Pings to those same servers still get through. Anything inside the network still works as expected.

After messing around bringing the bond interface on the router and computer up and down, unplugging and plugging in cables it eventually gets into a working state again. Each time it has happened it seems to need a different set of steps to get it unstuck from this state.

 

[KAD]

I can't say that my network is perfect, but I haven't noticed anything that I could point to a cause, or that LinkAgg has caused any issues on my network

although, I'm bonded to a switch that does both layer 2 and layer 3 routing
so I could forsee, that if there was some issue, the switch may resolve it before I notice

 

[ebalsumgo]

ok so i managed to get bonding working to an unmanaged switch. however i still run into the problem of all my wired connections dying... any suggestions?