From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Message-ID: <51AD0277.1060907@inti.gob.ar> Date: Mon, 03 Jun 2013 17:54:15 -0300 From: =?UTF-8?B?R2FicmllbCBUb2zDs24=?= MIME-Version: 1.0 References: <51A3C7E9.7080208@inti.gob.ar> <20130528070057.GC3333@ritirata.org> <51ACBCE4.8000506@inti.gob.ar> <20130603160813.GC1226@ritirata.org> In-Reply-To: <20130603160813.GC1226@ritirata.org> Content-Type: text/plain; charset="utf-8"; format="flowed" Content-Transfer-Encoding: 8bit Subject: Re: [B.A.T.M.A.N.] batctl bw performance Reply-To: The list for a Better Approach To Mobile Ad-hoc Networking List-Id: The list for a Better Approach To Mobile Ad-hoc Networking List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Antonio Quartulli Cc: The list for a Better Approach To Mobile Ad-hoc Networking Hi Antonio, El 03/06/13 13:08, Antonio Quartulli escribió: > Hello Gabriel, > > first of all thank you very much for testing the bw meter :-) > answers follow inline. You're welcome, I think it's an interesting idea. > > On Mon, Jun 03, 2013 at 12:57:24PM -0300, Gabriel Tolón wrote: >> Hi Antonio, >> >> As you suggested, I've tested the bandwidth meter with batman-adv-devel >> 38a1b72. I've installed it on two TP-Link WDR-3500 routers. >> > > oky. However, since the bw_meter has not been merged yet, its branch changes > from time to time because it gets rebased on top of the very latest development > code. Therefore, when you decide to update your testbed, I'd suggest you to look > at: http://git.open-mesh.org/batman-adv.git/shortlog/refs/heads/ordex/bw_meter > and pick the commit ID of the last patch at the top. > > If you have problems in doing this, please ask. I don't understand what commit ids should I use. For example, how could I find the ones you suggested me before for batman (38a1b72b6) and for batctl (eebba0bfb). Because on the top of the git I see for example the commit d672369c3476f002662fa7e83b76ee56a93a76db for "batman-adv: set maximum number of BW session", which seems in a different format, right? > >> Running normal iperf inside the routers, and without unloading any kernel >> module, I get 60 Mbps using UDP, 80 Mbps using TCP (I don't understand why >> TCP works better), and about 4 Mbps with batctl bw -t 10000. Any ideas? By >> the way, the batctl ping between routers doesn't work, I get time-out. >> Besides, there's another strange behaviour, when I try to pass traffic >> from a PC client connected to one router, to the other router using iperf, >> if I use UDP values of about 100 Mbps, the router that receives the >> traffic and redirects it, reboots. Thanks. >> > > Ok..well, first of all I'd rather check what is wrong in the mesh. If batctl > ping does not work, then it means that there is something else which is not > working/configured properly. Yes, it's weird, specially because there's connectivity between routers thanks to batman-adv. > > For this reason I'll first suggest you to update to this commit id: d672369 . > Then, does "batctl o" properly shows the other node with a "reasonable TQ"? So, in menuconfig I should select that id for batman-adv-devel right? I wanted to see that commit on the git page using the search field but I couldn't. By the way, how to know which batctl version use with each batman-adv-devel commit? Sorry if all these questions are too basic, but I've tried to compile with this d672369 version and I got some errors, so I'd like to understand if it was related to these things. > > > By the way, your iperf tests have been done on top of batman-adv? or directly on > the wifi interface? > > Cheers, > I got the results mentioned using interfaces managed by batman-adv. Regards Gabriel