My upstream ISP (my ISPs ISP) has stopped sending RAs before/when the initial RA expires.
The situation:
- wan interface (pppoe-wan) comes up
- DHCPv6 configuration happens
- an RA arrives with an 1800s (30m) lifespan
- default route is added to pppoe-wan
- (ip -6 route show match ::0)
- 1800 seconds pass, route is removed
- ipv6 networking stops routing to the outside world
This was working for a good while, and I’ve not changed my config, so I assume they stopped
sending timely RA updates a few days ago (when this problem started).
The RFCs don’t say they have to send updates - only that they can, and they should respond
to explicit solicitations with RAs.
I’ve worked around this with a script that checks for the default route every second or so and
issues an “ifup wan6” (which presumably triggers a DHCPv6 transaction and results in a fresh
RA (confirmed via tcpdump)).
But it seems to me that there should be a way to tell odhcp6 (I think that’s the right component,
since it looks like the v6 plumbing is handled in userspace) to send a solicitation at 50% or 75%
or whatever of the route’s lifetime (frankly not unreasonable for that to be the default behaviour).
Have I missed a config setting? Is there already a way to do this?