My ela cluster failed to load metadata when running, but I can't see the other problem. If there is the same problem, please help to take a look

Thanks for your reply, David, because I am only installing elasticserach now, the entire cluster of nodes are reporting this error, and the other nodes are also unhealthy. My storage is the remote storage OSS provided by the preparator. I think it should be that the cluster node file is overwritten, which means that maybe only one node's file is written, and the other two nodes' files are directly overwritten, resulting in constant reporting of missing files.