hi Peter, from later emails, it sounds like you have everything working. however, I do want to mention one alternative that does not require changing the controller.
if you add: --verbose=vconn:file:dbg --log-file="/dev/null" to the command to start ofprotocol in /lib/openflow/ofprotocol.sh, then the problem of the temporary disconnects also goes away. this logging overhead seems to be enough to re-jigger ofprotocol into correctly processing the ECHO_REPLY's coming from the controller in response to the ECHO_REQUEST's sent by the switch. as this is a band-aid, and not a solution, I once explored changing the frequency of the kernel timer, but this bricked my TPLINK as it prevented the ethernet device driver from loading. after re-flashing via a serial header, I didn't continue debugging, so that's all I know for certain. perhaps the logging is causing additional interrupts or time in the kernel? cheers, Andrew On Oct 17, 2012, at 2:47 PM, Peter Fales wrote: > Hi Yiannis, > > Thanks for the quick response to my mail. I've CCedd the mailing > list this time. > > To recap, we seem to be seeing the same problem on our WRTG54L with > Pantou that the was reported on a TPLINK router last June. And, > yes, like that previous question we are using POX as our controller. > > Does that mean the the problem is actually in the controller? Is > there a problem with the way the controller is handling the Echo Request? > The previous discussion indicated that it might be a timing issue, so > I tried artificially delaying the Echo Reply message, but that didn't > help. > > > On Wed, Oct 17, 2012 at 11:24:14AM -0700, Yiannis Yiakoumis wrote: >> Hi Peter, >> >> What controller do you use? Is this pox? >> Also, please consider adding the mailing list (openflow-discuss) for >> follow-ups.. >> >> Thanks, >> Yiannis >> >> On Wed, Oct 17, 2012 at 11:19 AM, Peter Fales < >> peter.fa...@alcatel-lucent.com> wrote: >> >>> Hi Yiannis, >>> >>> We've recently started experimenting with OpenFlow and have installed >>> the Pantou distribution on a WRT54GL. It seems to work (only some >>> basic sanity testing so far), but I think we are running into the >>> same issue described at >>> >>> https://mailman.stanford.edu/pipermail/openflow-discuss/2012-June/003367.html >>> >>> The symptoms appear to be exactly the same (except that the OP is using >>> a TPLINK router, and we're using Linksys). I see the original >>> question from last June, but I do not see any additional discussion since >>> then. Has any more information become available? >>> >>> Thanks in adnance for any suggestions you may have! >>> >>> -- >>> Peter Fales >>> Alcatel-Lucent >>> Member of Technical Staff >>> 1960 Lucent Lane >>> Room: 9H-505 >>> Naperville, IL 60566-7033 >>> Email: peter.fa...@alcatel-lucent.com >>> Phone: 630 979 8031 >>> > > -- > Peter Fales > Alcatel-Lucent > Member of Technical Staff > 1960 Lucent Lane > Room: 9H-505 > Naperville, IL 60566-7033 > Email: peter.fa...@alcatel-lucent.com > Phone: 630 979 8031 > _______________________________________________ > openflow-discuss mailing list > openflow-discuss@lists.stanford.edu > https://mailman.stanford.edu/mailman/listinfo/openflow-discuss _______________________________________________ openflow-discuss mailing list openflow-discuss@lists.stanford.edu https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
