Created attachment 1695539 [details] Logs Description of problem: When using an IPv6 provisioning network, the ironic fails to download the RHCOS image. The node PXE boots successfully and gets a valid IPv6 IP. After some minutes, the IP stops working, and ironic-python-agent fails (or just get stuck forever) as it couldn't reach the bootstrap node. Version-Release number of selected component (if applicable): Openshift 4.4 IPI How reproducible: Always, but the failure happens randomly on each of the nodes. One node may work, and the other two fail. Steps to Reproduce: 1. Configure install-config.yaml to use IPv6 as provisioning subnet (fd00:1101::0/64) 2. Run openshift-baremetal-install and open the server BMC console 3. Watch as the nodes fails or get stuck infinitely trying to download the image Actual results: ironic-python-agent fails to download the RHCOS image from the bootstrap node. Expected results: Image is downloaded and written to the disk successfully. Additional info: Ping from the bastion to any of the masters we are deploying: Machine PXE boots, loads the kernel and ramdisk, ping starts working: 64 bytes from fd00:1101::1fc: icmp_seq=14 ttl=64 time=0.269 ms 64 bytes from fd00:1101::1fc: icmp_seq=15 ttl=64 time=0.365 ms 64 bytes from fd00:1101::1fc: icmp_seq=16 ttl=64 time=0.339 ms 64 bytes from fd00:1101::1fc: icmp_seq=17 ttl=64 time=0.310 ms 64 bytes from fd00:1101::1fc: icmp_seq=18 ttl=64 time=0.267 ms 64 bytes from fd00:1101::1fc: icmp_seq=19 ttl=64 time=0.259 ms 64 bytes from fd00:1101::1fc: icmp_seq=20 ttl=64 time=0.261 ms 64 bytes from fd00:1101::1fc: icmp_seq=21 ttl=64 time=0.311 ms 64 bytes from fd00:1101::1fc: icmp_seq=22 ttl=64 time=0.330 ms From fd00:1101::2: icmp_seq=51 Destination unreachable: Address unreachable From fd00:1101::2: icmp_seq=52 Destination unreachable: Address unreachable From fd00:1101::2: icmp_seq=53 Destination unreachable: Address unreachable From fd00:1101::2: icmp_seq=54 Destination unreachable: Address unreachable From fd00:1101::2: icmp_seq=55 Destination unreachable: Address unreachable From fd00:1101::2: icmp_seq=56 Destination unreachable: Address unreachable From fd00:1101::2: icmp_seq=57 Destination unreachable: Address unreachable From fd00:1101::2: icmp_seq=58 Destination unreachable: Address unreachable From fd00:1101::2: icmp_seq=59 Destination unreachable: Address unreachable After a while it starts working again.
This is not a blocker for 4.5.0 GA. Setting target release to current development branch (4.6.0). For fixes (if any) requested/required on earlier versions, clones will be created for 4.5.z or earlier as appropriate.
This isn't a showstopper for 4.5.0 GA at this point. Setting target release to 4.6.0 (the current development branch). For fixes (if any) requested/required on prior versions, clones will be created targeting those z-stream releases as appropriate.
In our lab, there is no problem downloading the image over Ipv6. If there are special environment or routing we can simulate in the lab, please contact me on e-mail or slack
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (OpenShift Container Platform 4.6 GA Images), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:4196