#9347 Copr should be accessible via IPv6
Closed: Fixed 3 years ago by kevin. Opened 3 years ago by dominik.

Copr and its repos seems to be presently available only via IPv4.
They should be made available via IPv6.


@praiskup @msuchy Can you take this? Or do you need anything from our side to enable ipv6 on those instances?

For aws, we have had to enable dhcpv6 or assign a static ipv6 address, it doesn't seem to do SLLAC. @mobrien might have more info as he was dealing with that for proxies recently.

Metadata Update from @mohanboddu:
- Issue tagged with: medium-gain, medium-trouble, ops

3 years ago

Metadata Update from @mohanboddu:
- Issue priority set to: Waiting on Assignee (was: Needs Review)

3 years ago

Can you take this?

I think so. As long as we are in AWS, it should be doable I guess. There's
a very old bug with this request, but we used to be blocked by OpenStack:
https://bugzilla.redhat.com/show_bug.cgi?id=1232949

There's still is a long term plan to move copr infrastructure as much as
possible to OpenShift... but I believe we shouldn't loose ipv6 by that
move.

Or do you need anything from our side to enable ipv6 on those instances?

I don't know off-hand. We'll experiment on staging, and keep you updated.

@mobrien might have more info as he was dealing with that for proxies recently.

Yeah, any more info is welcome!

I recently did this for some servers running Fedora32 in AWS, I tried the DHCPv6 but it didn't work correctly for me.

This is the Amazon docs for DHCPv6 if you want to try it.
https://docs.aws.amazon.com/vpc/latest/userguide/vpc-migrate-ipv6.html#vpc-migrate-ipv6-dhcpv6

If that doesn't work to set static ipv6 for fedora assuming eth0 as the main network port.

edit /etc/sysconfic/network to ensure to set

NETWORKING_IPV6=yes

and edit /etc/sysconfig/network-scripts/ifcfg-eth0 to ensure the following are set

IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_FAILURE_FATAL=no
IPV6ADDR="<ipv6_address>"

A reboot of the instance should get ipv6 in action. ifconfig should show the new address under eth0.

This is ugly but will get you the ipv6 address on the instance if you want to automate it

curl -s http://169.254.169.254/latest/meta-data/network/interfaces/macs/$(curl -s http://169.254.169.254/latest/meta-data/mac)/ipv6s

AWS deployment has been always a temporal solution. During the period of datacenter migration and deployment of the new Copr server.
The New Copr server is deployed, but not functional yet.

@kevin you can assign this to me or @praiskup - I cannot change the metadata of tickets.

Metadata Update from @pingou:
- Issue assigned to msuchy

3 years ago

Did the changes to the Ipv6 in AWS and on site fix this?

Metadata Update from @smooge:
- Issue priority set to: Waiting on Reporter (was: Waiting on Assignee)

3 years ago

After some testing with devel copr instance, it should be OK. We'll try to assign IPv6 to production VMs tomorrow .... and then probably setup AAAA records. The negative is
that the IPv6 records aren't "elastic" so when we move to new set of VMs we'll have to update AAAA records.

We configured AAAA records now, so if it doesn't work it is likely because the records are not yet propagated?

ping6 doesn't work here, but curl works fine. :)

Can we close this now? Or is there more to do?

I think we can close this. Thanks!

Metadata Update from @kevin:
- Issue close_status updated to: Fixed
- Issue status updated to: Closed (was: Open)

3 years ago

Issue status updated to: Open (was: Closed)

3 years ago

Issue status updated to: Closed (was: Open)
Issue close_status updated to: Fixed

3 years ago

Login to comment on this ticket.

Metadata
Boards 1
ops Status: Done