Linux - NewbieThis Linux forum is for members that are new to Linux.
Just starting out and have a question?
If it is not in the man pages or the how-to's this is the place!
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
i need help with sorting out my ipv6 routing question. I find the whole IP thing really confusing and inclusion of IPv6 really doesn't help.
Here is what I got.
I got 3 /64 ipv6 blocks. A/64, B/64, and C/64. I have 3 nodes: a linux blade cluster, (LBC) a Quagga router and the switch.
The network looks something like this: LBC <--> Quagga <--> switch.
I have assigned A::5/128 to LBC as a virtual IP, meaning that as far as the rest of the world is concerned, LBC has A::5 while within the LBC, it can be whatever.
My goal is to have A::5/128 pingable from the switch.
Now for IPv4, to do something of this is actually rather easy. Like this:
a IPv4 vip would be defined, and attached to the blade cluster with a /30 subnet, and one side of the /30 would be on the LBC, while the other side would be on the Quagga router.
Then the quagga router would say, the /30, route through my default gw. which would be in the same subnet as the switch. say a /28.
The I just have to define, on the quagga, this IPv4 VIP should route through my /30, and on the switch my IPv4 VIP should route through my quagga's default gw.
After that, everything is pingable and i get a workable system.
but with IPv6, and the introduction of link local addresses, i am now confused as to what to route where.
in theory, i should take above example, replace ipv4 with ipv6, and it should work. but i am having a lot of problems with that. I know link local are valid only within their context, i have to ping from a specific network interface, unlike ipv4.
in theory, i should take above example, replace ipv4 with ipv6, and it should work.
Yes, that's exactly what you should do.
Quote:
Originally Posted by jzoudavy
but i am having a lot of problems with that.
Could you be a bit more specific?
Quote:
Originally Posted by jzoudavy
I know link local are valid only within their context, i have to ping from a specific network interface, unlike ipv4.
Link-local addresses are not relevant to routing issues, as they are just that: link-local. You can safely ignore them when you're configuring routing.
As long as you're not accessing the link-local addresses, IPv6 works the same way as IPv4. The main difference when it comes to routing is that every IPv6 router advertises itself on the network with ICMP6 Router Advertisement messages. Other hosts can pick that up automatically or query the network with ICMP6 Router Solicitation messages.
actually from the LBC to the Quagga it is done via ospf, so i need the link local. but that isn't the issue right now.
quagga and my switch seems to be having a fit. they are both under /64, but can't ping each other, even though the tcpdump shows the request and reply, and Neighbourhood Discovery on the switch shows the proper ips/mac and everything.
I'm not saying you don't need link-local addresses; after all, they are used as source addresses for most (all?) link-local multicast traffic. I'm just saying you don't have to include them in your configuration anywhere, as they are generated and used automatically.
If the switch and the Quagga router both have a statically assigned IPv6 address in the same /64 network and ND works as it should on both hosts, you should definitely be able to ping from one host to the other. I'd check the iptables rules and the /proc/sys/net/ipv6/conf/* settings on the Linux host, and check for ACLs that may exist on the switch.
I remember that an OSPF configuration error once caused a Quagga router to seemingly drop off the network, and I was unable to ping the router from other hosts on the same subnet/segment until the error was corrected. I've never been able to figure out exactly how that could happen, but you may want to stop the ospf6d daemon until you've verified that the IP configuration itself is working.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.