Hello All,
I am getting a problem with batman-adv-2013.0.0,
Please find the details,
I have two batman-adv nodes A and B,
Initial setup: A (bat0: 192.0.0.1) ---------- B (bat0: 192.0.0.2) Ping is working properly between node A and node B.
Setup after restarting B: A (bat0: 192.0.0.1) ---------- B (bat0: 192.0.0.2) After restarting node B, I am setting the same IP as previously to B (192.0.0.2) but this node (B) is not reachable from A for quite some time.
Note: If I assign some other IP to B after restart for example 192.0.0.3, node A is able to ping B without any delay.
Please let me know this is the problem due to DAT or something else.
Thanks Ajeet
On Tue, Mar 12, 2013 at 11:25:18 +0530, ajeet singh wrote:
Hello All,
I am getting a problem with batman-adv-2013.0.0,
Please find the details,
I have two batman-adv nodes A and B,
Initial setup: A (bat0: 192.0.0.1) ---------- B (bat0: 192.0.0.2) Ping is working properly between node A and node B.
Setup after restarting B: A (bat0: 192.0.0.1) ---------- B (bat0: 192.0.0.2) After restarting node B, I am setting the same IP as previously to B (192.0.0.2) but this node (B) is not reachable from A for quite some time.
Note: If I assign some other IP to B after restart for example 192.0.0.3, node A is able to ping B without any delay.
Please let me know this is the problem due to DAT or something else.
Hello Ajeet,
I thought about it and I think this is a consequence of DAT caching bat0's address. When you restart B, its bat0's MAC address is randomly renewed and therefore you have to wait for its IP address to timeout on A's DAT cache table.
If you change IP you don't see this problem because A does not have it in its cache and will store it along with the new bat0's MAC address.
Other than the bat0 mac address (which change at any restart) long timeouts are usually ok because the ARP entries do not change so often.
However, since batman-adv knows when a MAC address is removed from the network (see the Translation Table mechanism for this), we are thinking about a solution which combines the two components in order to proactively remove obsolete ARP entries from the DAT cache.
Thank you very much for reporting! If you have any better idea or suggestion feel free to share! :)
Cheers,
b.a.t.m.a.n@lists.open-mesh.org