> What did it try prior to 0A6E04? As in back when it was working? I'm not sure.
> If new ips, could redo nodeset (if boot to hard drive, nodeset sn1 boot) This fixed the problem. Although it already looked like the nodes were set to boot off the correct kernel image (here's a working and broken node): "cn35","xcat/osimage/centos7.1-x86_64-netboot-compute/kernel","xcat/osimage/centos7.1-x86_64-netboot-compute/initrd-stateless.gz","imgurl=http://10.110.7.105:80//install/netboot/centos7.1/x86_64/compute/rootimg.gz XCAT=!myipfn!:3001 NODE=cn35 FC=0 ifname=em1:34:17:EB:E9:53:EC netdev=em1 ",,,,, "sn1","xcat/osimage/centos7.1-x86_64-netboot-compute/kernel","xcat/osimage/centos7.1-x86_64-netboot-compute/initrd-stateless.gz","imgurl=http://10.110.7.105:80//install/netboot/centos7.1/x86_64/compute/rootimg.gz XCAT=!myipfn!:3001 NODE=sn1 FC=0 ifname=em1:b8:2a:72:fc:07:44 netdev=em1 ",,,,, ... I did a `nodeset sn1 boot` then set it back to the image. After this the nodes booted correctly. I have my shell history of when I renamed the nodes and it doesn't look like I did anything to tell xCAT to change the kernel image. So to me there appears to be a bug which caused this but I don't have time to track it down any further; issue resolved, on to the next one. Thanks for pointing me in the direction of a resolution. Jeff White HPC Systems Engineer Information Technology Services - WSU On 01/04/2016 12:58 PM, Jarrod Johnson wrote: > What did it try prior to 0A6E04? Does sn1 resolve to the same ip? If new > ips, could redo nodeset (if boot to hard drive, nodeset sn1 boot) > > -----Original Message----- > From: Jeff White [mailto:[email protected]] > Sent: Monday, January 04, 2016 3:08 PM > To: [email protected] > Subject: [xcat-user] Unable to boot nodes after renaming them > > I have xCAT 2.10 running on CentOS 7 with a number of stateless nodes. > This was working fine until I renamed these nodes: > > cn36 --> sn1 > cn37 --> sn2 > cn38 --> sn3 > cn39 --> sn4 > > I did that with `chdef` and also changed the nodes' IPs with `nodech`, I do > not know what those two tools did but in this case the nodes no longer boot. > They appear to be trying to use the wrong kernel image. > The error is: > > Trying to load: pxelinux.cfg/0A6E04 > Could not find kernel image: xcat/nbk.x86_64 > > The nodes which were not renamed continue to boot correctly. So: > > 1. What would chdef and nodech have changed to point these nodes to a > different kernel image? > 2. How do I set them back to the image every other node is using? > > > I noticed that under /tftpboot/pxelinux.cfg/ there were symlinks of the > old node names: > > lrwxrwxrwx 1 root root 4 Nov 23 11:33 0A6E0438 -> cn36 > lrwxrwxrwx 1 root root 4 Nov 23 11:34 0A6E0439 -> cn37 > lrwxrwxrwx 1 root root 4 Nov 23 11:35 0A6E043A -> cn38 > lrwxrwxrwx 1 root root 4 Nov 23 13:00 0A6E043B -> cn39 > > I manually changed this to the new names (mv cn36 sn1; rm 0A6E0438; ln > -s sn1 0A6E0438) but that appears to have no effect, the nodes still > fail with the same error. > ------------------------------------------------------------------------------ _______________________________________________ xCAT-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/xcat-user
