Our partner is trying to install 4.20-ec5 spoke cluster with ZTP/Gitops approach in disconnected environment. ACM version is 2.14 and MCE version is 2.9.0-212. When installation of master nodes start, master nodes completes the step "writing to disk" and then stuck at "configuring" step. When we check nodes from BMC console, we see that  there is a message of "ostree-prepare-root.service: Failed with result 'exit-code'. Failed to start OSTree Prepare OS/." and it keeps getting rebooted many times. It is not possible to ssh to these failing master nodes as well as they cannot be booted. We tried to restart installation of spoke cluster couple of times but each time one of the random master nodes fail at the same step.

Later on, we also tried to replicate same issue from OCP 4.18 hub with ACM 2.13 and MCE 2.8 for 4.20-ec5 spoke cluster and we had some similar "ostree" issue on some of the nodes as well. 

I am attaching the clusterinstance.yaml that is used during these spoke cluster deployment attempts.  Some screenshots from BMC console and some related logs.

Clusterinstance.yaml: https://drive.google.com/file/d/1vXbI1epzvstvW0wfmQd40-2w-DXPZ6ZB/view?usp=sharing 

Spoke cluster:
OCP 4.20-ec5
 
Hub cluster:
OCP 4.20-ec5
ACM 2.14
MCE 2.9.0-212

Deploy a OCP 4.20-ec5 spoke cluster with ZTP/Gitops approach on disconnected environment. Hub cluster can be either 4.18 or 4.20 and either with ACM 2.14 and ACM 2.13. 

Spoke Cluster deployment fails at step of "Configuring" on ACM console and BMC console of Nodes show "ostree" error.

Spoke cluster installation should be completed successfully.