On Friday, December 29, 2017 5:01:23 PM CET Robert Bates wrote:
Hello,
Is it possible to have b.a.t.m.a.n. ARP if packets are received at bat0 for a client which has been removed/deleted due to timeout, and is therefore no longer in the translation tables?
In one customer application of a product of ours (a mesh AP we've licensed from another vendor/developer, which is based on openWRT/b.a.t.m.a.n.), we are being adversely affected by the 10 minute inactivity timer on transtable_local. The clients in this customer's network/application are stationary devices which basically do not speak unless spoken to (e.g. when they are polled for data). They are periodically polled by a management platform, using an upper layer protocol running over TCP. The problem is that this customer's polling cycle time is variable, and occasionally it is taking longer than 10 minutes between successive polls of a given client/device. When this happens, that client is of course removed from transtable_local, and transtable_global on the other nodes in the mesh. Meanwhile, the polling/management platform has a very long ARP cache life, so it never ARPs (and apparently it is not possible on this platform to have the customer implement dynamic, rather than static ARP table entries, in which it would ARP upon polling failure). So once we get into this state, polls to this client device which has dropped out of the mesh are not possible, and their management platform throws alarms, etc. To bring it back in service at that point requires an ARP, which the customer is manually triggering with a ping, whenever one of these "outages" occurs.
We know that the transtable_local inactivity/removal timer value can be extended, and we will probably do that, but we would also like to know if it is possible to have b.a.t.m.a.n. ARP for the removed client in this case. We prefer this approach, rather than arbitrarily changing the tt_local timer to some value which may not work well in some other customer's network/application. I know that there is a statistically valid underlying assumption with this 10 minute inactivity timer on transtable_local, that clients will typically be "chatty". But again, that is not the case in this application, which is a very common one in the industry in which we operate, where clients are very often fixed devices which only respond to explicit queries or commands. This is a new product and protocol for us, and this could beg the question of whether or not b.a.t.man.-based meshing is the right solution in this type of application. We believe it can be; it would just be helpful if we can configure it to ARP in this type of scenario.
Can you please comment on how this might be possible (config or otherwise)?
Hi Robert,
batman-adv does not ARP on its own, so there is no way to configure this.
You should either increase the timer from 10 minutes to something (reasonably) high, or have another program sending some frames to refresh the ARP on behalf of the client - e.g. every 5 minutes, read the transtable local, send a packet for each MAC. However, in my opinion, this is just a more round-abot way with possibly more pitfalls on its own.
A general way to handle this would be to send those unicasts unknown to batman-adv as broadcasts. However, this would be problematic for networks with big broadcasts domains which already suffer from too high broadcast load, but have a sane ARP mechanism in place otherwise.
So long story short, increasing the timeout seems to be the most easy and effective solution to me.
Cheers, Simon