Managed Platform Unit Wiki

DOCUMENTATION

This is the MPU's technical documentation for computing staff. Anything labelled (MPU) is probably of interest mainly to the MPU.

BIOS

• MPUbiosUpgrade - how the BIOS was updated on a Dell PowerEdge server.
• SelectPCBiosSettings - BIOS settings for Select PCs
• ClonedBios - How to apply the standard BIOS configuration for a DICE machine using a USB stick.

Backups, Data, Disasters

• • CRITICAL SHUTDOWN - how to shut down servers in an EMERGENCY
  • MPU backups
  • MPU disaster recovery
  • General notes on how to restore from a backup
  • How to make a new LCFG master from backups
  • 2012 Flood Exercise
  • Replacing a failed disk on a Dell server which uses MegaCli/MegaRAID - including how to report the disk problem to Dell.
  • MPUGasExplosionJan2014
  • Full disk warnings on DICE machines
  • The Great AT Power Shutdown of January 2017

Booting

• • AutoReboot
  • How to force a reboot from a serial console
  • Booting Context - apply LCFG resources only while booting.
  • (MPU) DBan - how it was set up.
  • FreeDOS and desktop BIOS updates.
  • Tang and clevis - a short investigation of Network Bound Disk Encryption.

DICE LCFG Service

• • For more information about the vulnerabilities see:
  • wiki.lcfg.org is where to look for general LCFG information.
  • How to use dumpdeps - how to see LCFG dependencies - which hosts use which files, and vice versa.
  • How to stop the LCFG servers when you want to make several changes atomically. It also explains how to start them again...
  • (MPU) Important notes on managing the LCFG master and slaves
  • LCFG slave server status pages - lcfg1, lcfg2, lcfgtest
  • (MPU) LCFGMasterRecovery - How to make a new LCFG master from backups
  • Comparisons of LCFG component versions on supported platforms - http://lcfghost.inf.ed.ac.uk/cgi/component_versions.cgi
  • (MPU) Managing the main LCFG servers
  • (MPU) Moving the LCFG Master to new hardware
  • The test LCFG server
  • (MPU) Managing the LCFG website
  • (MPU) Upgrading the master LCFG server to SL5, to SL6
  • The inf layer - LCFG managed but without DICE add-ons, good for finding/eliminating DICE dependencies in the lcfg or ed levels so other Schools can use your stuff.
  • Managing Bugzilla
  • How to connect to the IS jabber server, which hosts the lcfg chatroom.
  • InterWikis lets us define handy wiki shortcuts for URLs. We have lots of local shortcuts for our LCFG repositories, bug trackers and so on. For example: Bug:123, Help:nx, LCFG:core/include/dice/options/small-server.h.
  • Future Platform Ideas - A gathering of ideas for future major platform upgrades.

Hardware and Firmware

• • How to report Dell hardware problems
  • DICE Disk Partitioning.
  • How to configure new monitors
  • Hardware testing ideas
  • Model-specific hardware information
    • Guidance on specifying a Dell server
  • CPU SpecInt scores
  • Information on the IBM DS3524 storage array
  • MegaCli Cheat Sheet (configures RAID drives on a PERC card) - and see also ServicesLocalDiskServer and another MegaCli cheat sheet
  • Replacing a failed disk on an R720 - although this is probably relevant to any machine using MegaCli/MegaRAID.
  • TooHot prevents server overheating
  • hwmon server hardware checks
  • For Dell server firmware updates use dsu or the Lifecycle Controller:
    • Dell System Update (dsu) is an easier way to update firmware on more recent Dell PowerEdge servers. (It replaces the old Server Firmware Updates system.)
    • You can also do Dell firmware updates using the Lifecycle Controller
  • (MPU) Upgrading or rebooting a KVM server - see also MPUKVMServerMultiUpgrade
  • iDRAC8 investigation
  • AsusFirmware - some attempts to update the firmware on Scan 3XS / ASUS ESC4000 G3S servers.
  • IntelMicroCode - information on Intel supported CPU microcode

Installs

• • LCFGLinuxInstall - explanation of the LCFG linux install process
  • PxeService - The DICE PXE Service
  • Install Contexts - apply LCFG resources only while installing.
  • SL6InstallErrors - a collection of error messages which we'd like to eliminate from SL6 installs.
  • (MPU) The LCFG install mechanism
  • (MPU) Building installroot

Inventory

• The Tartarus manual

Kernel

• Rebuilding the DICE kernel

Packages

• • Emergency Software Updates Howto
  • Which profile.packages resource to use
  • Local software packaging guidelines
  • How to find packages
  • How to build RPM packages using mock
  • (MPU) Package repository system
  • (MPU) Mirrors of various package repositories
  • (MPU) PatchingNss - how to concoct the local nss RPMs which make web browsers accept Ed Uni CA signed certificates
  • (MPU) How to restart the Package Forge incoming queue processor.
  • (MPU) Updating OS packages
  • (MPU) Updating LCFG to a new minor release
  • (MPU) Upgrading the Package Master
  • (MPU) Making package buckets by hand if you have to

Releases

• Release FAQ
• (MPU) Release Procedures

Virtualisation

• • Simple KVM service
    • KVM server scheduled downtime
    • Managing KVM guests
      • ... using kvmtool, rvirsh, virsh, virt-manager
      • Walkthrough of creating a KVM guest.
    • Managing a KVM host
      • Upgrading or rebooting a KVM server
  • (MPU) Virtual DICE management

Other

• • SwitchToSelfManaged - guidelines on how to switch a machine to being self-managed
  • Extra users and groups added with the auth component
  • AddingOptionsHeaders to configure optional features
  • Moving machine to a different wire without a reinstall.
  • WakeOnLAN - technical documentation for wake.inf
  • For CSOs: avoiding broken profiles when decommissioning etc. machines
  • Decommissioning Machines and Drives
  • MPUusefulLinks - useful web links
  • CabinAdmin - Admin notes for the log cabin service
  • Crackly Audio : In RT 64427 Henry Thompson pointed us at this fix for crackly audio. We haven't tried it ourselves.
  • SwitchDesk - Notes on the SL7 switchdesk application and how the dice desktop environment works.
  • DevelopDisasterMarch2016 - Notes on the "develop release disaster" of March 2016
  • MPUServersIPv6 - List of MPU servers which need IPv6 addresses
  • MPUSpectreMeltdown - Resources related to the Spectre Meltdown vulnerabilities
  • MPUManualEntitlements - Services, etc, with manually configured entitlements
  • XRDPService - XRDP remote desktop service
  • SSHService - Admin notes for the SSH service

Admin

• RT rota (from Monday 7th Oct 2013) - Chris - Monday, Tuesday, Wed am. Stephen - Wed pm, Thu, Fri
• From T1 2017 we've decided to aim to spend 20% of our time on personal development, 20% on operational work and 50% on project work.
• MPU meetings
• Current activities and To Do list
• MPUStarGazing - Areas we think we should be looking at (for staff development)
• MPU time monitoring
• Potential MPU talks
• Spending plans - 2019-22, 2018-21, 2016-19, 2015-18, 2014-17, 2013-16, 2012-15, 2011-14
• Review of computing.help pages

Projects

• MPU Projects - information on proposed, current and completed projects
  • Redevelop School Equipment inventory show, home-page,blog
  • Procedures for physical security - show home-page brain dump
  • Consider User Security Training Materials - show home
  • Review Security of LCFG Profile Access - show, home, blog
  • Investigate alternative DICE desktop platform - show, home, blog

MPU Servers

MPU servers' profile status on lcfg1, lcfg2. MPU servers' Nagios status.
The machines in server rooms have the rack/row numbers in their location to make it easier to find them.
See also ATServerRoomMainServerRacksPopulation, ForumRackPopulation, KBRackPopulation.

Name	Model	Role	OS	Location	Purchased	Notes
canopus	Dell R240	LCFG master	sl7	IF-6/36	2019 Jun
lute	Dell R440	General access XRDP	sl7	IF-5/16	2019 Mar
archlute	Dell R440	General access XRDP	sl7	AT-1	2019 Mar
clavichord	Dell R740	KVM server	sl7	KB-1/11	2019 Jun
harpsichord	Dell R740	KVM server	sl7	KB-1/8	2019 Jun
regulus	Dell R230	RPM cache/PXE	sl7	IF-6/24	2018 Apr
maia	Dell R230	RPM cache/PXE	sl7	AT-1/23	2018 Apr
banjo	Dell R740	KVM server	sl7	AT-0/5	2018 Jun
mandolin	Dell R740	KVM server	sl7	AT-2/2	2018 Jun
deneb	Dell R330	Packages mirror server and local package master	sl7	IF-1/34	2017 Mar
altair	Dell R230	LCFG slave server (lcfg3)	sl7	IF-6/34	2017 Mar
vega	Dell R230	LCFG slave server (lcfg4)	sl7	AT-1/25	2017 Mar
gaivota	Dell R730	KVM server	sl7	IF-3/14	2016 Jun
girassol	Dell R730	KVM server	sl7	IF-4/16	2016 Jun
salamanca	Dell R320	Offsite disaster recovery	sl7	KB-3/29	2015 April
azul	Dell R730	KVM server	sl7	IF-5/18	2015 Jun
bruegel	Dell R210 II	student.ssh	sl7	AT-2/21	2014 May
steen	Dell R210 II	Staff SSH	sl7	IF-11/21	2014 May
vermelha	Dell R720	Distance learning XRDP	sl7	KB-3/17	2014 Jun
amarela	Dell R720	KVM server (user VMs)	sl7	KB-3/35	2014 Jun
waterloo	Dell R720	staff XRDP	sl7	IF-0/8	2013 Apr
oyster	Dell R720	KVM server (test/devel)	sl7	AT-1/16	2013 Jun
jubilee	Dell R720	Spare	sl7	IF-4/20	2012 Jun
juice	HP DL180 G6	openafs build host	sl7	AT-1/20	2011 Jun

Virtual Machines

Desktop Machines

You can edit this data at MPuMachineList.

Old Stuff

• ClonedBios - How to apply the standard BIOS configuration for a DICE machine using a USB stick.
• (MPU) How the Server Firmware Report works
• (MPU) How to download new server firmware updates
• VMware deprecation progress
• MPUServersSL5 - Prioritized list of MPU servers for upgrading to SL5
• MpuRescue - for rescuing abandoned documentation
• MPUTransitionList - List of work that MPU members need to drop and the MPU unit needs to absorb
• RT Rota - MPU weekly RT Duty rota
• MPU bugzilla searches: LCFG, Linux, Solaris, Others. (Details.)
• MPUOurCalendar - MPU Calendar
• Old DICE LCFG Information
• MPUSolarisResources
• MPUBootstrapNewFedora - quick guide to how to get a new Fedora running
• MPUFcFiveInfSixtyFourBit - Notes on developing and managing fc5 64 bit machines
• MPUInfFC3 - Notes on managing "inf" level FC3 machines
• FC6 Upgrade
• FC5 Upgrade
• FC3 Upgrade
• Managing Solaris Patches
• DICEonMacPro - how to install DICE on a Mac Pro
• On hardware RAID headers
• DICE on VMware on DICE - how to run virtual DICE machines on a real DICE machine using VMware.
• Where MPU data is located on our servers
• VirtualisationComparison - Alastair's notes on his experiments with various virtualisation technologies.
• VMware: how to access a guest console, create a new guest, create a new host, move guests to another server, fixes, Choosing a technology
• (MPU) Adding a new network bridge to a running VMware server, (Cached copy)
• (MPU) Managing the PXE installer (SL5 only)
• LCFG Level Package Lists for SL5.
• SL5 Server Upgrade
• How to connect to the old lcfg chatroom on IS's old sidra.ucs jabber server.