Linux upload bonding using teql: Difference between revisions
Tony Hoyle (talk | contribs) No edit summary |
Tony Hoyle (talk | contribs) No edit summary |
||
Line 1: | Line 1: | ||
This is a simpler method of upload bonding than others, where you just want to round robin packets to the upstream, which is the normal case for AAISP. |
This is a simpler method of upload bonding than others, where you just want to round robin packets to the upstream, which is the normal case for AAISP. |
||
⚫ | |||
⚫ | |||
⚫ | |||
In /etc/ppp/ip-up.d place the following script: |
In /etc/ppp/ip-up.d place the following script: |
||
Line 19: | Line 13: | ||
ip -6 route replace default scope global dev teql0 |
ip -6 route replace default scope global dev teql0 |
||
</pre> |
</pre> |
||
⚫ | |||
⚫ | |||
⚫ | |||
[[Category:3rd Party Routers]] |
[[Category:3rd Party Routers]] |
Revision as of 15:51, 3 April 2017
This is a simpler method of upload bonding than others, where you just want to round robin packets to the upstream, which is the normal case for AAISP.
In /etc/ppp/ip-up.d place the following script:
#!/bin/sh sysctl -w net.ipv4.conf.ppp0.rp_filter=2 sysctl -w net.ipv4.conf.ppp1.rp_filter=2 tc qdisc add dev ppp0 root teql0 tc qdisc add dev ppp1 root teql0 ip link set teql0 up ip route replace default scope global dev teql0 ip -6 route replace default scope global dev teql0
You may wish to add ppp3, etc. as pppd sometimes allocates a different number overriding the unit directive. There's no harm in having too many.
The sysctl is so that the system doesn't get upset at packets going out of one interface and coming in another. I'm not certain this is required, but all examples seems to have it.
The teql device does not need an IP address, as it's virtual. It seems to behave in a sane way wrt. originating IP in all cases I've tried.