Hello, So here is where I believe things stand with my nodes that won’t complete setup (WCS), when compared with the nodes that will complete setup CS).
* Both CS and WCS nodes are stateless * Both CS and WCS nodes use the same image (CentOS 7) * WCS nodes were in the same group as CS nodes (though some of them have been moved for testing) * Tried WCS nodes without postbootscripts (though CS nodes used postbootscripts) * WCS nodes will boot into the Genesis kernel * CS nodes have been rebuilt many times, in an effort to compare differences and ensure that it wasn’t a corrupt boot image * WCS nodes end up hanging with the error “dracut-initqueue[417]: Read error (Connection timed out) in headers.” * CS nodes are only in rack 1, while WCS are in other racks * According to my rack person, the switch configs are the same for rack 1 and the other racks (which I believe, since I can boot into the Genesis kernel on any node) * All WCS nodes worked when being used with another xCAT master (both WCS and CS nodes have recently been moved to a test environment that is to mirror production) Any thoughts or ideas?
_______________________________________________ xCAT-user mailing list xCAT-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xcat-user
