Mailing List Archive

OSPF wrong configuration on mininet
Hello, I'm using OSPF on mininet and I'm having some issue
.[cid:part1.49BBF8C8.DE1A87C6@live.com]
I'm setting the link speed with mininet when defining the topology, however this doesn't seem to affect the ospf cost computation.
By sniffing some traffic I've noticed that all the LS Update messages contain a metric=100, even though the different link speed should affect the computation.
For example to go from R6 to R4 I'd expect a cost of 2 going through Ri4 and a cost of 20 going through R5.
This is an example of the ospf.conf file I'm using on each router:

hostname r6

log file /var/log/quagga/r6.log

router ospf
auto-cost reference-bandwidth 100
network 172.168.6.0/24 area 0
network 172.168.34.0/24 area 0
network 172.168.5.0/24 area 0
network 172.168.35.0/24 area 0

Could the area parameter be causing this unexpected behavior?
Re: OSPF wrong configuration on mininet [ In reply to ]
Hello Alessandro,

Could you provide additional configuration? Maybe one of the routers full
config except passwords.
Also, in a scenario with 1Gbps links, I would set the reference-bandwidth
at 1000. That is not your issue though.

Best Regards,

Ralph Mondegar


On Tue, Dec 19, 2017 at 9:31 PM, Alessandro Gaballo <alessandrogbl@live.com>
wrote:

> Hello, I'm using OSPF on mininet and I'm having some issue
> .
> I'm setting the link speed with mininet when defining the topology,
> however this doesn't seem to affect the ospf cost computation.
> By sniffing some traffic I've noticed that all the LS Update messages
> contain a metric=100, even though the different link speed should affect
> the computation.
> For example to go from R6 to R4 I'd expect a cost of 2 going through Ri4
> and a cost of 20 going through R5.
> This is an example of the ospf.conf file I'm using on each router:
>
> *hostname r6*
>
> *log file /var/log/quagga/r6.log*
>
> *router ospf*
> *auto-cost reference-bandwidth 100*
> * network 172.168.6.0/24 <http://172.168.6.0/24> area 0*
> * network 172.168.34.0/24 <http://172.168.34.0/24> area 0*
> * network 172.168.5.0/24 <http://172.168.5.0/24> area 0*
> * network 172.168.35.0/24 <http://172.168.35.0/24> area 0*
>
> Could the area parameter be causing this unexpected behavior?
>
> _______________________________________________
> Quagga-users mailing list
> Quagga-users@lists.quagga.net
> https://lists.quagga.net/mailman/listinfo/quagga-users
>
>
Re: OSPF wrong configuration on mininet [ In reply to ]
Hi Ralph, attached you find a .tar.gz containing all the configuration file I'm using in my topology.
Concerning the link speed, I'm using mininet TCLink and I set their speed when I create the topology.

I'm really struggling with this problem and I need to fix it ASAP, so tell me if you need anything else.

On 19/12/2017 19:01, Ralph Mondegar wrote:
Hello Alessandro,

Could you provide additional configuration? Maybe one of the routers full config except passwords.
Also, in a scenario with 1Gbps links, I would set the reference-bandwidth at 1000. That is not your issue though.

Best Regards,

Ralph Mondegar


On Tue, Dec 19, 2017 at 9:31 PM, Alessandro Gaballo <alessandrogbl@live.com<mailto:alessandrogbl@live.com>> wrote:
Hello, I'm using OSPF on mininet and I'm having some issue
.[X]
I'm setting the link speed with mininet when defining the topology, however this doesn't seem to affect the ospf cost computation.
By sniffing some traffic I've noticed that all the LS Update messages contain a metric=100, even though the different link speed should affect the computation.
For example to go from R6 to R4 I'd expect a cost of 2 going through Ri4 and a cost of 20 going through R5.
This is an example of the ospf.conf file I'm using on each router:

hostname r6

log file /var/log/quagga/r6.log

router ospf
auto-cost reference-bandwidth 100
network 172.168.6.0/24<http://172.168.6.0/24> area 0
network 172.168.34.0/24<http://172.168.34.0/24> area 0
network 172.168.5.0/24<http://172.168.5.0/24> area 0
network 172.168.35.0/24<http://172.168.35.0/24> area 0

Could the area parameter be causing this unexpected behavior?

_______________________________________________
Quagga-users mailing list
Quagga-users@lists.quagga.net<mailto:Quagga-users@lists.quagga.net>
https://lists.quagga.net/mailman/listinfo/quagga-users
Re: OSPF wrong configuration on mininet [ In reply to ]
Hello Alessandro,

I haven't gone through all your files but I've noticed that your
zebra.conf files are empty.
You must configure zebra with basic configuration, like IP Address of
each interface, speed, duplex.

Best Regards,

Ralph Mondegar

On Tue, Dec 19, 2017 at 11:39 PM, Alessandro Gaballo <alessandrogbl@live.com
> wrote:

> Hi Ralph, attached you find a .tar.gz containing all the configuration
> file I'm using in my topology.
> Concerning the link speed, I'm using mininet TCLink and I set their speed
> when I create the topology.
>
> I'm really struggling with this problem and I need to fix it ASAP, so tell
> me if you need anything else.
>
>
> On 19/12/2017 19:01, Ralph Mondegar wrote:
>
> Hello Alessandro,
>
> Could you provide additional configuration? Maybe one of the routers
> full config except passwords.
> Also, in a scenario with 1Gbps links, I would set the
> reference-bandwidth at 1000. That is not your issue though.
>
> Best Regards,
>
> Ralph Mondegar
>
>
> On Tue, Dec 19, 2017 at 9:31 PM, Alessandro Gaballo <
> alessandrogbl@live.com> wrote:
>
>> Hello, I'm using OSPF on mininet and I'm having some issue
>> .
>> I'm setting the link speed with mininet when defining the topology,
>> however this doesn't seem to affect the ospf cost computation.
>> By sniffing some traffic I've noticed that all the LS Update messages
>> contain a metric=100, even though the different link speed should affect
>> the computation.
>> For example to go from R6 to R4 I'd expect a cost of 2 going through Ri4
>> and a cost of 20 going through R5.
>> This is an example of the ospf.conf file I'm using on each router:
>>
>> *hostname r6*
>>
>> *log file /var/log/quagga/r6.log*
>>
>> *router ospf*
>> *auto-cost reference-bandwidth 100*
>> * network 172.168.6.0/24 <http://172.168.6.0/24> area 0*
>> * network 172.168.34.0/24 <http://172.168.34.0/24> area 0*
>> * network 172.168.5.0/24 <http://172.168.5.0/24> area 0*
>> * network 172.168.35.0/24 <http://172.168.35.0/24> area 0*
>>
>> Could the area parameter be causing this unexpected behavior?
>>
>> _______________________________________________
>> Quagga-users mailing list
>> Quagga-users@lists.quagga.net
>> https://lists.quagga.net/mailman/listinfo/quagga-users
>>
>>
>
>