Unfortunately, it appears xCAT breaks when trying to create a diskless image. That's probably another issue with 18.04.5. I think we'll just install 18.04.2 on the head node, and repeat all our configuration steps.
Calvin Dodge On Wed, Mar 17, 2021 at 8:30 AM Nathan A Besaw <bes...@us.ibm.com> wrote: > > Have you made any progress on this problem or are you still stuck? > > Can you deploy the node with a diskless image based on the same OS and see > the disks from within the booted OS? > > The newest version of Ubuntu 18 that xCAT officially supports is 18.04.2. > Several xCAT users have moved forward to 18.04.5, but > there appear to be some issues with xCAT and Ubuntu 18.04.5 that are not well > understood. > > I don't have access to this combination of hardware and OS, perhaps other > members of the list have some experience with this combination. > > Calvin Dodge ---03/12/2021 06:41:04 PM---We are trying to deploy C6420 nodes > with a diskfull image. The deployment hangs about about 9.5 sec > > From: Calvin Dodge <caldo...@gmail.com> > To: xcat-user@lists.sourceforge.net > Date: 03/12/2021 06:41 PM > Subject: [EXTERNAL] [xcat-user] Problem deploying Dell C6420 from Ubuntu > 18.04.05 > > ________________________________ > > > > We are trying to deploy C6420 nodes with a diskfull image. The > deployment hangs about about 9.5 seconds after the kernel recognizes > the network and USB devices. > > Installment details: Head node was deployed using the live server ISO > ubuntu-18.04.5-live-server-amd64.iso. > The xcat-go install process created the /install/OS folder using that > ISO. But we could not nodeset a node with that osimage name until we > ran copycds manually with the regular server ISO > ubuntu-18.04.5-server-amd64.iso. > > We've seen other strangenesses, like xCAT looking for a folder named > "ubuntu-" when we tried to run genimage to create a diskless image > > Meanwhile, the nodes are using PERC controllers for storage, which > uses the megaraid_sas kernel module. When we unpack the initrd used > in the initial PXE load, we don't find the kernel module there. Could > that be the source of our deployment hang? We see instructions on > adding modules, but they don't appear to be relevant, because (1) Dell > doesn't seem to provide a driver disk for the initial kernel > (5.4.0-22), and (2) the "use RPM" approach is unlikely to work with a > distribution that doesn't use RPMs. > > We CAN install Ubuntu on the node using the server ISO, so the > necessary drivers are present there, at least. So it seems to be an > xCAT issue. > > Has anyone else encountered this issue? If not, how can we diagnose > it, beyond adding (an as yet unfound) megaraid_sas kernel module to > the initial PXE initrd? > > Calvin Dodge > [attachment "cat_node_install_hung[33146].PNG" deleted by Nathan A > Besaw/Poughkeepsie/IBM] _______________________________________________ > xCAT-user mailing list > xCAT-user@lists.sourceforge.net > https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.sourceforge.net_lists_listinfo_xcat-2Duser&d=DwICAg&c=jf_iaSHvJObTbx-siA1ZOg&r=95pZYYPXXV-5mn9YO7FX6q2JdbPYeCR6fMnFTgqZ_M8&m=jOGxgBpk2RY0sFkf4czZC_mW22LlbQeGFkAqdSoeE-M&s=aq3MGS4v8yB6srQiaWjx2rj79VGt3k7500bcUFnxfXw&e= > > > _______________________________________________ xCAT-user mailing list xCAT-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xcat-user
