[prev in list] [next in list] [prev in thread] [next in thread] 

List:       quagga-dev
Subject:    [quagga-dev 6756] Re: [PATCH] Fix config file processing so result
From:       paul () jakma ! org
Date:       2009-06-30 17:47:21
Message-ID: alpine.LFD.2.00.0906301838360.13450 () stoner ! jakma ! org
[Download RAW message or body]

On Tue, 30 Jun 2009, Chris Caputo wrote:

> If you are suggesting a "neighbor <address> peer-group xxx" command 
> be in the global section rather than AF section, that results in 
> peer_group_bind() returning BGP_ERR_PEER_GROUP_AF_UNCONFIGURED and 
> displaying the message:
>
>  "Activate the peer-group for the address family first"

It needs to be in both sections - as per the inline diff :).

>>> You'll also have silently lost "neighbor 2001:db8::1 password
>>> testpass" because the "neighbor 2001:db8::1 peer-group v6peer" after
>>> it in the config, will reset the peer's config.
>>
>> Because you havn't set them for v6.
>
> The parser won't let you enter "password" and "passive" lines, among
> others, in the "address-family ipv6" node.  So if they are received while
> processing a config file (or vty for that matter), they will eject the
> flow out of the "address-family ipv6" node to be read by the "router bgp
> xxx" node.

Ok, so we need to fix the peer-group config-copy thing to not clobber 
global state, then?

>> - consider settings in what is currently the global+IPv4 section to be
>>   global and make peer_group2peer_config_copy inherit them
>>
>> ?
>
> Intriguing.  I don't know.  I'm hesitant to deviate from Cisco's layout,
> if that is what we are matching.

Well, a sane write-out format would be good. If we need a 2-pass 
parser for what we write out now, something is wrong with what we 
write out, no?

regards,
-- 
Paul Jakma	paul@jakma.org	Key ID: 64A2FF6A
Fortune:
Pity the meek, for they shall inherit the earth.
 		-- Don Marquis
_______________________________________________
Quagga-dev mailing list
Quagga-dev@lists.quagga.net
http://lists.quagga.net/mailman/listinfo/quagga-dev
[prev in list] [next in list] [prev in thread] [next in thread]