> > >-----Original Message----- >From: [email protected] [mailto:sipx-dev->[email protected]] On Behalf Of Lawrence, Scott (BL60:9D30) >Sent: Tuesday, April 14, 2009 2:29 PM >To: Krzeminski, Damian (BL60:9D30) >Cc: [email protected] >Subject: Re: [sipX-dev] should sipXconfig prevent adding redundant proxies >if it thinks your network is not up for it? > >On Tue, 2009-04-14 at 13:22 -0400, Damian Krzeminski wrote: > >> I need some input here: >> http://track.sipfoundry.org/browse/XCF-3616 > >> XCF-3616 suggests that if admin selected host name as a SIP domain name >we >> should stop them from adding redundant servers. Do people think we should >> really be that restrictive? It is conceivable that one prefers to >> reconfigure DNS to make it work to changing SIP domain, right? > >Technically, it actually _should_ be possible to make it work to make >one of the hosts and the domain have the same name. You _could_ >configure SRV records that map host1.example.com to both >host1.example.com and host2.example.com. In theory, a SIP UA should >figure that out. But there's one of my favorite reminders: > > In theory, there's no difference between theory and practice; > in practice, there is. > >I wouldn't bet that the majority of implementations wouldn't be confused >by hitting the same name as an SRV and an A record (I _think_ that >sipXtackLib would do the right thing, but not having tested it, one >never knows). > >> As far as I know SIP domain that is equal of your host name is not even >the >> default configuration. So if someone went through the troubles of >changing >> this default once why should we stop them from changing it again. > >Just because they changed it doesn't mean they knew what they were >doing, or what it would cost them later. > >> Just to remind where I am coming from. sipXconfig is not a tool for >> policing your network. If you happen to have invalid hostnames or strange >> DNS configured sipXconfig can point out the errors but it really should >not >> prevent you from configuring something. > >But remember who will have to answer the questions when what we allow >doesn't work... > >I would use the rule that we should restrict configurations to a subset >of the theoretical possibilities that we believe is easily doable, will >work well with most implementations, and still has the flexibility to do >things that actually need to be done. > >I think that in this case that rule would suggest that if someone wants >to add a server, they first be told that they must reconfigure the >domain name so that it is not identical to the first host name. >
Nicely put. The whole purpose of DNS/DHCP testing is to GUIDE the admin towards a working config. It does not help if an expert can get it to work. What we need to do is assist a less skilled installer to get it to work without having to call for help. --martin _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
