This document describes a series of steps to validate xCAT configuration prior to configuring Moab to manage hardware via xCAT. It is assumed the reader is familiar with xCAT and the xCAT configuration on the target site. This document does not provide xCAT configuration documentation or troubleshooting information; please refer to the xCAT documentation for such information.
Verify that all nodes that Moab will manage are known to xCAT with the xCAT nodels command. Ensure that all expected (and no unexpected) nodes are listed. You may find it useful to create new group names to identify Moab managed nodes.
[root@h0 moab]# nodels hyper,compute h1 h2 h3 h4 h5 h7 kvmm1 kvmm10 kvmm2 kvmm3 kvmm4 kvmm5 kvmm6 kvmm7 kvmm8 [root@h0 moab]#
Verify that all nodes report their status correctly using the xCAT nodestat command. Ensure that all nodes show the correct status (sshd, installing, noping, and so forth); there should not be any timeouts or error messages.
[root@h0 moab]# nodestat hyper,compute |sort h1: pbs,sshd h2: pbs,sshd h3: pbs,sshd h4: pbs,sshd h5: pbs,sshd h7: noping kvmm10: noping kvmm1: pbs,sshd kvmm2: pbs,sshd kvmm3: pbs,sshd kvmm4: pbs,sshd kvmm5: pbs,sshd kvmm6: pbs,sshd kvmm7: pbs,sshd kvmm8: noping kvmm9: noping [root@h0 moab]#
Verify that all nodes that Moab will manage have hardware management interfaces correctly configured using the xCAT nodels and rpower commands. After each of the rpower commands, verify the requested state was achieved with rpower stat.
[root@h0 moab]# nodels h1,kvmm1 nodehm.mgt nodehm.power h1: nodehm.power: ilo h1: nodehm.mgt: ilo kvmm1: nodehm.power: kvm kvmm1: nodehm.mgt: kvm [root@h0 moab]# rpower h1,kvmm1 off h1: off kvmm1: off [root@h0 moab]# rpower h1,kvmm1 stat h1: off kvmm1: off [root@h0 moab]# rpower h1,kvmm1 boot h1: on reset kvmm1: on reset [root@h0 moab]# rpower h1,kvmm1 stat h1: on kvmm1: on [root@h0 moab]#
Verify that all operating system images that Moab uses are configured correctly in xCAT. For stateful images, test that all combinations of operating system, architecture, and profile install correctly.
[root@h0 moab]# rinstall -o centos5.3 -a x86_64 -p hyper h1 h1: install centos3.2-x86_64-hyper h1: on reset [root@n100 ~]# sleep 15 && nodestat n05 n05: ping install centos5.3-x86_64-hyper [root@h0 moab]#
For stateless images, test that nodes are able to network boot the images.
[root@h0 moab]# nodech h5 nodetype.os=centos5.3 nodetype.arch=x86_64 nodetype.profile=hyper [root@h0 moab]# nodeset h5 netboot h5: netboot centos5.3-x86_64-hyper [root@h0 moab]# rpower h5 boot h5: on reset [root@h0 moab]# sleep 60 && nodestat h5 h5: pbs, sshd [root@h0 moab]#
If you use VM migration, verify that xCAT can successfully perform migrations using the rmigrate command.
[root@h0 moab]# rmigrate kvmm7 h1 kvmm7: migrated to h1 [root@h0 moab]# ssh h1 virsh list Id Name State ---------------------------------- 33 kvmm1 running 34 kvmm2 running 35 kvmm7 running