minidesign_xcatprobe

Overview

We want to make a new command to probe all the possible issues in xCAT. It can probe xCAT MN and xCAT node definition statically. It also can probe the node discovery and node deployment staticaly. The goal is to make a command to help xCAT users to predict and debug xCAT problems easily.

Interface

The syntax of the xcatprobe command:

xcatprobe <probe_type> [parameters]

• xcatprobe # same as xcatprobe help
• xcatprobe help
• xcatprobe nodedef <noderange>
• xcatprobe osdef <osimage>
• xcatprobe xcatmn
• xcatprobe DB
• xcatprobe node <noderange>
• xcatprobe switch
• xcatprobe nodeready <noderange> [console] [deployment]
• xcatprobe nodediscover
• xcatprobe nodedeploy

xcatprobe help

Display the usage of xcatprobe.

• Display the basic usage
• Display all the probe type

xcatprobe nodedef

Probe node definition.

• Check validate of node name
• Check ip<=>node entry in /etc/hosts
• Check DNS resolution
• Check HWcontrol: check definition and try the rpower status to make sure hwcontrol is ready for using.
• Check attributes: mgt, netboot, mac

xcatprobe osdef

Probe the definition of OSimage

• check the basic attribtues: imagetype, osarch, osdistroname, osname, osvers
• check the existence of packages in pkgdir
• check the packages in the otherpkgdir
• check the entries in the pkglist and otherpkglist
• check the rootimage in rootimgdir for netboot image

xcatprobe xcatmn

Probe the readiness of xCAT MN

• check he hostname, long name
• check the selinux has been disabled
• Probe the xCAT has been installed successfully
• Check xcatd has been started sucessfully
• check the basic configuration of xcat: site table, passwd table, network table
• Check the network services are running configured properly: dhcpd, named, tftpd, httpd
• Check the size of dhcpd.leases file less than 100MB
• check the disk space of /tmp /var /install

4. xcatd,

xcatprobe nodeready [console] [deployment]

Probe the readiness of node.

• Check the console configuration

  • check the node attributes: cons, serial*
  • check the cfg in /etc/conserver.cf

• Check the readiness for OS deployment:

  • provmethod is set, readiness of osimage
  • dhcp set in dhcpd.leases
  • readiness of bootloader and bootloader cfg file
  • readiness of installer kernel + initrd
  • readiness of installer cfg file

• It can handle all the nodes in the