Changes to rs2.ape.net.nz - 10am Tomorrow
Hola. We're in the process of final testing prior to livening up [animating?] a new route server to take the place of rs2.ape.net.nz on new hardware. There will be a brief outage tomorrow as we swap the servers. Note: rs1.ape.net.nz will not be affected by this change. The new server is dual-stacked and will support IPv6 on 2001:7fa:4:c0cb::9a02/64 - all the current v6 peers are configured on this server, feel free to configure up a session to the new server. Any issues, please let us know.
Hola.
I'm so popular I have to reply to my own email.
There will be a brief outage tomorrow as we swap the servers.
They've been swapped. We're seeing the expected number of routes on the v4 route server.
The new server is dual-stacked and will support IPv6 on 2001:7fa:4:c0cb::9a02/64 - all the current v6 peers are configured on this server, feel free to configure up a session to the new server.
We're seeing some sessions to the new v6 route server, feel free to add more. Please let us know if you're seeing any issues. At this stage we're planning on swapping in the dual-stacked replacement for rs1 and rs1-v6 this coming Thursday 2nd August at 10am. To make everyones lives [perhaps] a little easier the new server will have the following addresses configured 192.203.154.1 2001:7fa:4:c0cb::9a01 2001:7fa:4:c0cb::9a03 - the old rs1-v6 address. FYI: we anticipate rolling out similar changes in Wellington within the next 3 weeks I'll update with confirmed times and dates.
At this stage we're planning on swapping in the dual-stacked replacement for rs1 and rs1-v6 this coming Thursday 2nd August at 10am.
2 things. Firstly - a reminder that we're making this change tomorrow at 10am. Secondly - on this new build of the route servers we're not importing routes learnt from the APE peers into the kernel routing table on the route server, this has the effect of making the route servers unreachable from anything other than directly connected APE hosts, this is by design. A few participants have noticed this as they were apparently monitoring connectivity via ICMP. Please adjust any monitoring accordingly.
participants (1)
-
Sid Jones