Yes, you can use a formula as well. I don't use that feature because I am using Ansible to auto-generate the configuration; I leave the computation to Ansible rather than xCAT.
One of the drawbacks of using a formula in xCAT is that it doesn't actually do IP math; it performs math on each decimal in the dotted-decimal notation. As a result, you will run into trouble if your subnets aren't /24. For instance, if you want to have 500 nodes and want to use a /23 to hold them, you will run into problems with using an xCAT formula. Yes, I have multiple names in DNS for each node - actually, four. I also have a different name for the BMC, and for one additional network. Don't worry about the sysconfig files - that's what xCAT does for you. I believe it's the hardeths postscript that configures all the NICs, but am not sure off the top of my head. hostname returns the "main" node name, without any of the suffixes (but as an FQDN) And for your most recent question: I just checked - the NIC that is used for PXE booting continues to get its address from DHCP. The other NICs have statically assigned IPs. There is no /etc/hostname, I assume that one of the postscripts calls the hostname command to set it. All this applies to stateless nodes, and I believe most of this is configurable. _______________________________________________________________________ Kevin Keane | Systems Architect | University of San Diego ITS | kke...@sandiego.edu Maher Hall, 192 |5998 Alcalá Park | San Diego, CA 92110-2492 | 619.260.6859 *REMEMBER! **No one from IT at USD will ever ask to confirm or supply your password*. These messages are an attempt to steal your username and password. Please do not reply to, click the links within, or open the attachments of these messages. Delete them! On Wed, Feb 13, 2019 at 9:24 AM Thomas HUMMEL <thomas.hum...@pasteur.fr> wrote: > On 2/13/19 3:33 PM, Thomas HUMMEL wrote: > > >> I'm using the ip and nicips.eth0 and nicips.eth2 attributes on the > >> node to configure the correct IP addresses that I want. Also, be sure > >> to use nicextraparams.eth2=MTU=9000 if you want Jumbo frames on your > >> 10 GB network. > > > > Ok, I'll check that out. > > I've just read the man. > Your idea is interesting indeed, even in the case you wouldn't mind > mixing administrative and workload data in the 10Gb/s connection as it > would remove the fear of PXE on the "wrong" nic. > > The only thing is that we use a formula in the hosts table for the > primary ip (actually the only ip we have for one node, BMC ip aside of > course) so it's derived from the node number/name. Would nicpips > attribute support the formula ? > > >> Finally, I'm using the nichostnamesuffixes attribute so the nodes have > >> different, but predictable, names on the 1 Gb network and the 10 Gb > >> network. > > So you've got 2 names in DNS after makedns I guess ? > But how does it translate in sysconfig scripts/files (or the equivalent > is non centos/redhat OS) on the node ? What does 'hostname' returns if > run on the node ? > > Thanks. > > -- > Thomas H. > > > > _______________________________________________ > xCAT-user mailing list > xCAT-user@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/xcat-user >
_______________________________________________ xCAT-user mailing list xCAT-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xcat-user
