[Olsr-dev] Feedback on current OL

David Murray (spam-protected)
Mon Jun 9 12:57:14 CEST 2008


Quoting Aaron Kaplan <(spam-protected)>:

>>
> did you observe over a longer period of time?
>
> My experience is that it takes a long time until ir really drops to  
> 1.0 (or slightly above)
> due to the new exponential backoff.

Yep, It has been about 30 mins now.

>
>
>> This new machine is getting a metric of 1.061 over the Ethernet link
>> instead of 1.150 like the suspect machine. Just a question, why is the
>> metric 1.061 instead of 1.000. I think that it is unlikely that an
>> uncongested Ethernet link would drop a packet.
>>
>>>
>>> /hannes
>>>
>>> David Murray wrote:
>>>> Quoting Hannes Gredler <(spam-protected)>:
>>>>
>>>>> david,
>>>>>
>>>>> did you see UDP checksum errors in outgoing or incoming packets ?
>>>>
>>>> They are outgoing hellos. The interface is an Ethernet Interface.
>>>>>
>>>>> /hannes
>>>>>
>>>>> David Murray wrote:
>>>>>> Hi guys, just a bit of feedback on the latest development branch
>>>>>> found here [1]. I have had it running in my test setup (8 dual
>>>>>> radio ALIX nodes) for about 3-4 days. It seems to be working
>>>>>> great, however, one problem I have found is with the the metric.
>>>>>>
>>>>>> Currently, I have a 100mbit Ethernet link that is being given a
>>>>>> metric of ~1.150. Iperf shows that this link provides ~100 mbit
>>>>>> which means that there is nothing physically wrong with the link.
>>>>>> To investigate a little further I pulled up wireshark and took a
>>>>>> look at the OLSR hellos.
>>>>>>
>>>>>> The wireshark output was showing some UDP checksum errors "maybe
>>>>>> caused by UDP checkum offload" on some olsr hello packets. Perhaps
>>>>>> if anyone else is using this branch they could confirm whether
>>>>>> they are also getting the errors or if it is something specific
>>>>>> with my equipment.
>>>>>>
>>>>>> Thanks
>>>>>> Dave
>>>>>>
>>>>>> [1] http://gredler.at/hg/olsrd/archive/tip.tar.gz
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>>
>>
>> -- 
>> Olsr-dev mailing list
>> (spam-protected)
>> http://lists.olsr.org/mailman/listinfo/olsr-dev
>
> ---
> there's no place like 127.0.0.1
>
>
>






More information about the Olsr-dev mailing list