Hi Shrutarhi, Destiny is definitely the recommended branch, that's right.
just like you noticed, monitoring generates xids itself, and the epoch is specific to it. The epoch is just some variable (secs) defined on the top of the file. However it also increments the xids when an asynchronous user request occurs (e.g. when a user double-clicks on a switch on the gui to get it's current flow table). So the xid is incremented for each periodical or manual request. If I recall correctly, an earlier patch on destiny generalized this so that all requests (from any component) have their xid automatically generated and filled by NOX's packing library (unless you override them and manually assign one). This is definitely true for C++ components, but I think it has been exposed to the python side too (relevant patches<http://noxrepo.org/cgi-bin/gitweb.cgi?p=nox&a=search&h=refs%2Fheads%2Fdestiny&st=commit&s=+xid+> ) On Fri, Dec 9, 2011 at 10:31 AM, Shrutarshi Basu <ba...@cs.cornell.edu>wrote: > Hi Kyriakos, > I've been looking through the monitoring app and figuring out how to use > xids. It seems like you're using collection_epoch to generate new xids and > collection_epoch is just a logical clock. However I couldn't tell when you > are advancing the clock. Is this application specific or is there some rule > for generating xids that I'm missing? > Thanks, > Basu > > > On Fri, Dec 9, 2011 at 12:44 PM, Shrutarshi Basu <ba...@cs.cornell.edu>wrote: > >> Thank you, I just checked out the Destiny branch. From what I can tell >> Destiny seems to be the recommended branch for NOX development. Is that >> correct? >> >> >> On Thu, Dec 8, 2011 at 7:47 PM, Kyriakos Zarifis >> <kyr.zari...@gmail.com>wrote: >> >>> Ah, you're also looking at coreapps/examples/monitor.py >>> I was referring to netapps/monitoring.py (you can find that in the >>> destiny branch) >>> (monitor.py is a much simpler, early monitoring component) >>> >>> >>> On Thu, Dec 8, 2011 at 10:26 AM, Shrutarshi Basu >>> <ba...@cs.cornell.edu>wrote: >>> >>>> Thanks, it looks like I am working with an older version of NOX -- I >>>> can see a monitor.py app that does not seem to use the xid at all. I will >>>> try to upgrade and see what I can do. Will report back. >>>> Thanks, >>>> Basu >>>> >>>> >>>> On Wed, Dec 7, 2011 at 4:27 PM, Kyriakos Zarifis <kyr.zari...@gmail.com >>>> > wrote: >>>> >>>>> Hi Shrutarshi, >>>>> >>>>> the field you should be looking at is the xid (transaction id). >>>>> I don't remember the specifics right now - I think in Destiny it is >>>>> automatically set (and returned?) and you can also set it when you send >>>>> the >>>>> request (that's what the monitoring module does) >>>>> >>>>> I can be more specific when I look at the code to refresh my memory, >>>>> but this might help for now >>>>> >>>>> On Wed, Dec 7, 2011 at 12:29 PM, Shrutarshi Basu <ba...@cs.cornell.edu >>>>> > wrote: >>>>> >>>>>> If I send a bunch of aggregate stats requests for different flows, is >>>>>> there a way for the aggregate_stats_in handler to know which response it >>>>>> is >>>>>> being called on? I see that it gets the dpid, but what if I have multiple >>>>>> flows on the same dpid that I want information on? >>>>>> Thanks, >>>>>> Basu >>>>>> >>>>>> -- >>>>>> -- >>>>>> Shrutarshi Basu >>>>>> Basus.me >>>>>> The ByteBaker <http://bytebaker.com> -- because Computer Science is >>>>>> not about computers >>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> nox-dev mailing list >>>>>> nox-dev@noxrepo.org >>>>>> http://noxrepo.org/mailman/listinfo/nox-dev >>>>>> >>>>>> >>>>> >>>> >>>> >>>> -- >>>> -- >>>> Shrutarshi Basu >>>> Basus.me >>>> The ByteBaker <http://bytebaker.com> -- because Computer Science is >>>> not about computers >>>> >>>> >>> >> >> >> -- >> -- >> Shrutarshi Basu >> Basus.me >> The ByteBaker <http://bytebaker.com> -- because Computer Science is not >> about computers >> >> > > > -- > -- > Shrutarshi Basu > Basus.me > The ByteBaker <http://bytebaker.com> -- because Computer Science is not > about computers > >
_______________________________________________ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev
