Otherwise, we risk to lose peer list when any failure happens after
the creation but before storing the list and as a consequence,
the weave container gets restarted.
Added log lines for packets that we will broadcast;
Added log lines for flows we don't know how to handle;
Moved the line for when we expect to insert a flow down to the miss-handler;
Packets handled by the bridge already have their own logging via --pktdebug
The tracker's HandleUpdate method is called by the ring update routines.
In addition, we add a dummy overlay for AWSVPC to make `weave (status|
report)` to produce reasonable reports which includes AWS hosts
connection state.
The nil FlowOp check in collectFlowOps was not addressed with the
introduction of DiscardingFlowOp. This meant that DiscardingFlowOps
produced in the router core would get handled by fastdp as if they were
foreign FlowOps, preventing flow creation and so leading to unwarranted
ODP misses.
To fix this, collectFlowOps cannot simply eliminate FlowOps for which
Discards() is true, because it's true for odpFlowKeyFlowOp. So the
Discards() case has to be handled in FastDatapath send().
Reworked interface to router so we split setting up the initial list
from beginning to make connections - we shouldn't do that until
after router.Start().
The current minimum PMTU we can detect is eight bytes, which, being
lower than the minimum IPv4 MTU of 68 bytes, can cause an exception in
the sleeve packet fragmentation implementation. We increase that minimum
to 552, the lowest PMTU value that Linux will cache.
The kernel processes vport deletion asynchronously, meaning that we can
sometimes fail to recreate vxlan vports on startup with an EADDRINUSE
error. We compensate by retrying a small number of times with a short
sleep on encountering this error, failing as before on a different error
or exhaustion of retries.
If overlay encryption is required between peers, the fast datapath
forwarder returns an error rather than aborting the process, allowing
the overlay switch to fall back gracefully to an encrypted sleeve
overlay connection.
Forward can return nil if the remote peer has HasShortID false. That
can happen if we learn about a 1.2 peer from a 1.1 peer and attempt to
connect to it.
This change just fixes the panic. HasShortID will never be set, so we'll
fall back to sleeve. But if that occurs, if can be fixed by a restart of
the affected routers.
This is mostly a matter of moving some files, changing package names,
changing some imports, and changing some type references.
We duplicate the odd constant & global var, and the intmac/macint
functions, since neither 'router' or 'mesh' nor, really,' common' are
a natural home for them.
Martynas Pumputis