Custer Tools Files and Scripts

• Documentation, File and scripts for the installation and operation of a new cluster
• Some end user tools

Topical References

• ClusterToolsCommon
  • The files common_functions, common_vars, common_host and NODES are the primary configuration and support scripts for all cluster utilities
• ClusterToolsDNSMASQ
  • PXE BOOT, DNS, DHCP services
• ClusterToolsPXE
  • PXE boot configuration files and setup
  • See also PXEBootFilesExample - TFTPBOOT example
• ClusterToolsFirewall
  • Firewall NAT documenation
• ClusterToolsIPMITOOL
  • Tools to reset, reboot into PXE boot once, reboot into BIOS setup once, power on power off, update user access, etc tools
  • ipmireset is a very useful tool for getting a hung node to reset
• ClusterToolsIDRAC
  • Used only on Dell Servers
• ClusterToolsImaging
  • Node imaging tools
• CUDA install utils and tests
  • ClusterToolsCUDA
• ClusterToolsSH2TWIKI
  • Much of the documentation was extracted from scripts automatically using markup

C Programs

• basestr.c
  • :Match substring
  • Returns: return( strncmp(argv[1],argv[2],len) == 0 ? 0 : 1);
• dhcp.c
  • DHCP server tester - wrapper for nagios plugin
• dnstest.c
  • Test how gethostname, gethostbyname, getdomainname , works vs how Legator Networker resolves hostnames.
  • This is especially an issue when a host had more then one name.
• fixname.c
  • Escape strings so they can be safely passed to a shell
  • Names may be quoted with ''s
• lc.c
  • University of Waterloo List Catalogue command
• sh2twiki.c
  • Convert BASH and C markup to HTML or TWIKI
  • See ClusterToolsSH2TWIKI
  • Much of the documentation was extracted from scripts automatically using markup
• testfree.c
  • Usage: testfree device -v -f threshold(in Megabytes)

Scripts and file summary

• accounts - default cscf-adm and cscf-op encrypted passwords
• cleanup - Delete files under a directory DIR and children that are older the DAYS old
  • ./cleanup directory days
• common_functions, common_vars, common_host and NODES
  • Common shell scripts used by most scripts under /cscf-adm/src
• cscf-root-pub-key.txt
  • copy of CSCF root public key to be added to /root/.ssh/autorized_keys2 file
• distroy_and_clone
  • himrod specific scripit added to RIPimage
  • Partitions and formats all drives on a node - mounts /dev/sda1 on /mnt/sda1
  • Runs: rsync -a -x -H --numeric-ids homrod:/himrod.node/ /*

• fetch_clonezilla - grabs lates clonezilla image and installes it in /tftpboot/pxes/clonezilla
  • see: /tftpboot/pxes/pxelinux.cfg/clonezilla_menu
• fetch_pxe
  • Capture existing PXE code and images for the Image Deployment System
  • Not used on himrod
• fetch_rip
  • Downloads lates version of RIP
• fetch_ubuntu_installers
  • Downloads Ubuntu PXE instalers for any iso images under /tftpboot/pxes/rawiso/Ubuntu
  • Updates /tftpboot/pxes/ubuntu-installer/menu.conf called from master pxelinux menu
• fetch_ubuntu_releases
  • Downloads ubuntu iso images to /tftpboot/pxes/rawiso/Ubuntu
  • Example: ./fetch_ubuntu_releases 12.04 14.04 beta*
• firewall * NAT firewall service for system
  • Uses /usr/local/bin/common_host and helper scripts common_vars and common_functions
• fixhost
  • change hostname and ssh keys for a host - update postfix
• fixperms
  • Update the permissions on files under a directory
  • Only update those files that have them wrong so that backups do not update files needlessly
• flush_cache
  • This script will cause the kernel to drop clean caches, dentries and inodes from memory, causing that memory to become free.
• format_disks
  • Partition and Format disks on a himrod node to prepare for imaging
• grub_fix
  • Updates /etc/default/grub, purges then reinstalls grub
• install_admin_users
  • Add local Administrative and Samba users defined in $ADM_USERS - see common_host
• install_dell_utils
  • Install Dell OpenManage Ubuntu Repository - not used on himrod
• install-networker
  • Install Legato Networker for CSCF backup services
• install_packages
  • Install common Ubuntu packages required for the Image Deployment System
• install_pxe
  • Install the PXE code and images for the Image Deployment System
  • We only install install the files once from the pxe folder
  • We should consider merging updates FROM the pxe folder to the pxe folder or the reverse
  • Correctly we use fetch_pxe (this excludes files that can be regenerated)
  • make fetch_pxe will do this
• install_samba
  • setup cscf-adm and cscf-op local user
    • Prompts to create an iinitial samba password for each account
• install_syslinux
  • Install and compile syslinux (the default Ubuntu version is out of date)
  • Installs files in /tftpboot/pxes/syslinux
• install_ubuntu_scripts
  • not used on himrod
• IPMIview
  • Supermicro IPMI utility wrapper
  • Console access and IPMI settings
• lc.c and lc
  • UofW list catalog program
• list_samba_users
  • List current samba users
• Makefile
  • Master make file calls all install and update scripts
• mntpart
  • used to chroot mount a partition - used by recovery tools like RIP
• mntproc
  • bind mounts partitions into chrooted environment
• NODES
  • Defines all of the cluster nodes by name and ILOM name
• pxe2usb.sh
  • Converts himrod PXE boot enviornment into standalone USB KEY
• pxe_liveboot_nfs
  • Service to mount Live Boot images
• readiso
  • Read an CD or DVD and create an ISO image
• sh2twiki and sh2twiki.c
  • Creates TWIKI and HTML documenation form simplified Markup Language
• syslinux-6.02.tar.gz
  • Syslinux source
• testfree and testfree.c
  • Used to verfify a partition is mounted and has enough free space
• test_vars
  • Lists variables used by common_vars, common_host and common_functions
• update_fstab
  • Updates /etc/fstab on nodes to include NFS home directories
• update_rip
  • Apply local costomizations to local RIP images
• updates
  • Add additional repositories, NTP and SMTP configuration, created symlinks need for CS AD
  • Used on clean system after first install
• wakeup uses wake on lan feature to wakeup any host listed in /etc/dnsmaqsq.hosts