[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: address pair exploration, flooding and state loss



Iljitsch van Beijnum wrote:
On 27-mei-2005, at 22:57, Erik Nordmark wrote:

2. The box reboots in 30 seconds. TCP retransmissions for the next 10 minutes are silently dropped because the retransmissions arrive at the peer's TCP with a bad checksum (due to "missing" shim6 rewrite).
Thus any RST doesn't arrive until 10 minutes or so later!


Can we create a new ICMP message that says "I got a packet that may or may not have come from you and it had a wrong layer 4 checksum"?

And modify every transport protocol implementation? I'd prefer if we don't have to rely on such modifications.


On a slightly related note: it would be good if we can rebuild shim state after it has been lost on one side when at the same time, there is a failure. This would allow new sessions to be created and immediately rehomed that would otherwise have failed.

Presumably when A discovers that B has lost the shim state and it had failed over to an alternate locator pair, it will immediately try to rebuilt it, before sending any more ULP packets.
In this case it makes sense for A to use the locator pair that was working before the state loss for the shim6 signaling (and if possible, tell B which locator pair was used in the reverse direction).


I don't know how important it would be to optimize this case though.

  Erik