I'm trying to install an OCP 4.8.2 bare metal cluster using the assisted installer. The discovery and install prep went fine, but then the install got stuck; the nodes are all "waiting for ignition" or otherwise failing to contact the machine config service on port 22623. Looking at the logs for the machine config service, I see: I0809 19:41:38.194289 1 bootstrap.go:37] Version: v4.8.0-202107011817.p0.git.29813c8.assembly.stream-dirty (29813c845a4a3ee8e6856713c585aca834e0bf1e) I0809 19:41:38.194406 1 api.go:65] Launching server on :22624 I0809 19:41:38.194433 1 api.go:65] Launching server on :22623 I0809 19:47:13.693042 1 api.go:110] Pool master requested by address:"128.52.62.137:42374" User-Agent:"Ignition/2.9.0" Accept-Header: "application/vnd.coreos.ignition+json;version=3.2.0, */*;q=0.1" I0809 19:47:13.693079 1 bootstrap_server.go:66] reading file "/etc/mcs/bootstrap/machine-pools/master.yaml" I0809 19:47:13.694091 1 bootstrap_server.go:86] reading file "/etc/mcs/bootstrap/machine-configs/rendered-master-7cc4ff6b0c0903a34703c4c86d9ff87e.yaml" I0809 19:47:18.098980 1 api.go:110] Pool worker requested by address:"128.52.62.244:59586" User-Agent:"Ignition/2.9.0" Accept-Header: "application/vnd.coreos.ignition+json;version=3.2.0, */*;q=0.1" E0809 19:47:18.099059 1 api.go:129] couldn't get config for req: {worker 0xc00060e200}, error: refusing to serve bootstrap configuration to pool "worker" I0809 19:47:18.116060 1 api.go:110] Pool worker requested by address:"128.52.62.134:60782" User-Agent:"Ignition/2.9.0" Accept-Header: "application/vnd.coreos.ignition+json;version=3.2.0, */*;q=0.1" And then that final error message just repeats forever. I am going to try aborting and restarting the install, so I won't be able to service requests for additional information from the cluster.
After resetting the install and trying again, we're getting a very similar error: E0813 20:01:08.704977 1 api.go:129] couldn't get config for req: {worker 0xc00063f280}, error: refusing to serve bootstrap configuration to pool "worker" I0813 20:01:08.757313 1 api.go:110] Pool worker requested by address:"128.52.62.231:42004" User-Agent:"Ignition/2.9.0" Accept-Header: "application/vnd.coreos.ignition+json;version=3.2.0, */*;q=0.1"
Hi Lars, Thank you for reporting this. Can you please share your cluster-id so we can take a closer look?
@lars this is expected behavior. Workers will not join on bootsrap control plane.
Igor, What are you talking about? This results in all the nodes getting stuck in the "waiting for ignition" state and failing to install. This is absolutely not "expected behavior" unless we expect the assisted installer to always fail and be completely useless as a product. I'm happy to hop on a call to look at this issue, but the bug is blocking a major project and should be fixed rather than closed.
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 500 days