Glad it works. Interesting, looking at it now it still looks like
OFPFW_ALL==0x3fffff should be working too. Something to look into, thanks
for bringing it up.

On Mon, Feb 21, 2011 at 5:35 PM, Malveeka Tewari <[email protected]>wrote:

> Setting flow.wildcards  = 0xffffffff solved the problem :-)
>
> Thanks!
>
> ----- Original Message -----
> From: "Kyriakos Zarifis" <[email protected]>
> To: [email protected]
> Cc: [email protected], "Murphy McCauley" <[email protected]>
> Sent: Monday, February 21, 2011 5:01:52 PM GMT -08:00 US/Canada Pacific
> Subject: Re: [nox-dev] Getting individual flow stats for an python app
>
> So the thing is that flow.wildcard definitely needs to be at least 0x3fffff
> (and not 0xffff, which I am guessing you got from the monitor.py example,
> which is actually obsolete and should be updated, as the wildcard field has
> been changed to 32 bits (actual max length is 22).
>
> So OFPFW_ALL (which should be equal to 0x3fffff, or 4194303) should cover
> it.
>
> As a fast check, could you try manually setting flow.wildcard = 0xffffffff
> just to make sure that it's a wildcard issue?
>
>
> On Mon, Feb 21, 2011 at 2:58 PM, <[email protected]> wrote:
>
>> Even with flow.wildcard = openflow.OFPFW_ALL the counters seem to be zero
>> .. :(
>>
>>
>> ----- Original Message -----
>> From: "Kyriakos Zarifis" <[email protected]>
>> To: [email protected]
>> Cc: [email protected], "Murphy McCauley" <[email protected]>
>> Sent: Monday, February 21, 2011 2:47:10 PM GMT -08:00 US/Canada Pacific
>> Subject: Re: [nox-dev] Getting individual flow stats for an python app
>>
>> can you try using "flow.wildcards = openflow.OFPFW_ALL"
>>
>> after making sure that nox.lib.openflow has been imported?
>>
>>
>>
>> On Mon, Feb 21, 2011 at 1:37 PM, <[email protected]> wrote:
>>
>>> Thanks!
>>> I have another question.
>>>
>>> Using the nox controller , I have inserted forwarding rules such that
>>> packets that match a particular (src_mac, dst_mac) pair  be sent out on a
>>> particular port.
>>> Since this is a wildcard entry, it also matches packets that belong to
>>> the null vlan as long as the (src_mac, dst_mac) pair matches.
>>>
>>> However, when I request for aggregate flow stats from the nox controller,
>>> it does not show any flows
>>>
>>> 00117|openflow-event|DBG:received stats reply from 00ffca7e3ad4
>>> Aggregate stats in from datapath 00.:ff:ca:7e:3a:d4
>>> {'packet_count': 0L, 'byte_count': 0L, 'flow_count': 0L}
>>>
>>> This is how I am requesting for aggregate flow stats
>>>  def aggregate_timer(self, dpid):
>>>         flow = ofp_match()
>>>         flow.wildcards = 0xffff
>>>         self.ctxt.send_aggregate_stats_request(dpid, flow,  0xff)
>>>         self.post_callback(MONITOR_TABLE_PERIOD, lambda :
>>> self.aggregate_timer(dpid))
>>>
>>> But running ovs-ofctl on the switch does show that there are flows in the
>>> forwarding table.
>>>
>>> Jan 02 11:25:34|00001|ofctl|INFO|connecting to
>>> unix://var/run/openvswitch/ful0.mgmt
>>> stats_reply (xid=0xe9c1a): flags=none type=2(aggregate)
>>>  packet_count=414 byte_count=42228 flow_count=2
>>>
>>> I am not sure why sending a request from the nox controller does not give
>>> the correct response. I don't know if that's a bug with the switch's
>>> implementation or nox or I am doing something wrong while requesting for
>>> stats.
>>> The following thread leads me to believe that I am missing something but
>>> I am not sure what.
>>>
>>> https://mailman.stanford.edu/pipermail/openflow-spec/2010-February/000890.html
>>>
>>> Any thoughts ?
>>>
>>> Thanks!
>>> Malveeka
>>>
>>>
>>>
>>>
>>> ----- Original Message -----
>>> From: "Kyriakos Zarifis" <[email protected]>
>>> To: "Murphy McCauley" <[email protected]>
>>> Cc: "Malveeka Tewari" <[email protected]>, [email protected]
>>> Sent: Monday, February 21, 2011 1:10:42 PM GMT -08:00 US/Canada Pacific
>>> Subject: Re: [nox-dev] Getting individual flow stats for an python app
>>>
>>> Exactly - there's an example of using this in netapps/monitoring.py
>>>  (line 188) in destiny, (and you can trace how it's handled etc)
>>>
>>> If all you need is to print out/look at the flow stats (rather than act
>>> on them programatically), and you do switch to destiny, you could use the
>>> gui for that.
>>>
>>> On Mon, Feb 21, 2011 at 8:44 AM, Murphy McCauley <[email protected]> wrote:
>>>
>>>> I believe that the flow stats event is Pythonized in both the zaku and
>>>> destiny branches -- it's just that a helper method to make the callback 
>>>> easy
>>>> to register is missing from zaku.  A simple register_for_flow_stats_in
>>>> method DOES exist in the destiny branch, though.  So I'd suggest you either
>>>> switch to the destiny branch, or look at the relevant patch to destiny
>>>> (21e3bf59aca5d3df8e86bcd09cc9156a66866f7f) which should demonstrate how to
>>>> register a callback for flow stats events (it's not particularly hard --
>>>> basically you call register_handler with
>>>> Flow_stats_in_event.static_get_name() as the first parameter).
>>>>
>>>> Hope that helps.
>>>>
>>>> -- Murphy
>>>>
>>>> On Feb 21, 2011, at 8:30 AM, Malveeka Tewari wrote:
>>>>
>>>> Hi
>>>>
>>>> I am writing a python app for the NOX  controller and want to obtain
>>>> flow statistics for individual flows from an openflow switch.
>>>> I can see that nox supports python callback handlers for handling stats
>>>> reply for  aggregate flows stats, table stats , port stats and desc stats
>>>> events, but the handler for individual flow stat seems to be missing.
>>>>
>>>> Any suggestions on how can I extend the functionality to handle flow
>>>> stats event in a python nox app.
>>>>
>>>> Thanks!
>>>> Malveeka
>>>> _______________________________________________
>>>> nox-dev mailing list
>>>> [email protected]
>>>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> nox-dev mailing list
>>>> [email protected]
>>>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
>>>>
>>>>
>>>
>>
>
_______________________________________________
nox-dev mailing list
[email protected]
http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org

Reply via email to