Review Project List and Prioritise

There was a long discussion on the current and pending projects for the next project prioritisation round.

  • LCFG Server comes back for T2 but dependent on Simon's availability
  • RHEL 6 based port (131) - started already
  • Investigating Replacement for VMware server is done, and leads into a new project...
  • Implement Virtualisation service - needs new project entry (see below)
  • SL6 LCFG port aims to be finished by end of April (still to do - MPU config at DICE layer, LCFG server level stuff (fibre-channel/bonding))
  • SL6 - project for upgrading MPU machines (count as a development project - not operational) - New project
  • AFS Automation - AFS - Chris has some time in T1 for this but it may go into T2
  • Repository Mirroring (140) - already marked down for T2 but likely to be below line if we do Server refactoring
  • Virtualized DICE images - scoping (probably would go RaT unit for Iain R)
  • Base Dice ??
  • Install scripts for self-managed machines : yes, but probably T3
  • Boot component rewrite - no, wait for upstream change in technology
  • Enhanced fstab component - could do a wee project for 2TB disk fix, leave more major rewrite for now
  • LCFG documentation technology - need to improve external take-up but no time this project round
  • Logserver improvements go to Inf unit
  • LCFG component test framework - leave until after refactoring client project
  • LCFG client refactor - needs to wait until after server work
  • F14 port to inf layer - don't do this until next year by which time it'll be F16, if needed
  • Consider Puppet et al- Add a new project to review/evaluate other 'configuration management tools', especially if integrated into upstream vendor distributions
  • Rewrite syslog component
  • Automatic upstream
  • Package Submission - enhancements - remember buckets
  • Remote hardware management for Desktops - sleep/wake-on-lan etc - merge with Power Saving project and rename

We discussed the implentation details for a local virtualisation service based on KVM. This needs an SL6 server so will need to be postponed slightly into T2, but for now some minimum documentation on how to setup the service could be produced eg. how to use libvirt. The question was raised if the virtualisation service should be run by MPU or if each unit could have it's own virtualisation service. This then raises the issue of ease of moving VMs. in the event of machines being down etc

Projects

Software Build Farm

The problem from last meeting was still ongoing. Some code had been reworked. The pkg builders will reach stable release this week.

SL6

64 bit - done. Full install - not tested. Core LCFG packages - done. Stephen is working on OpenAFS, but SL6 does the 'wrong thing' in setting up AFS for 1.4.14, patches required. PXE booting - needs native SL6. We need to document the SL6 INF layer, and put up a couple of VMs for testing (on KB server).

AFS Automation Project

Chris has spoken to Craig and Services unit have a wiki page with the proposed AFS requirements. Top priority is a script to quickly promote ReadOnly volumes to ReadWrite in an emergency. Stephen would show Chris what's involved in setting up a database server + file server for AFS.

Submit project

Chris was enthusiastically testing after finding out about Perl test scripts.

Misc Dev

patches for updaterpms - mail bug, - reboot request , run components after pkg updates, post normal pkg install/ post reboot request didn't create new schema

LCFG Slave Performance testing from the procurement point of view , what is server is best to go for. Keeping the whole file system in shared memory could bring build times down to 30 minutes. Otherwise disk is an issue and CPU too. DL180s won't do so we need to scope out a cheap server.

Operational

Fibre Channel devices - some aren't coming up fast enough on boot so get missed - telford slow at scanning local disks, more investigation required

Machine reboots - still six to do - cochin, rendall, schiff, split, tobermory and trondra

Activities List

New kernel - done

8 gig - cache requirement for AFS clients - after issues with installing RaT software

Test LCFG - done

Condor - dropped - by operational meeting. CSG should be consulted to confirm teaching/research requirement.

Moving VMs to KB on activities list.

DR stuff at KB - Add to list

Time management buckets - for T2 T3, CEG looking at better categoristion - user driven/event driven/routine maintenance

Newsletter -> becomes systems blog - add to agenda as a fixed item Chris suggested some topics

This Week

  • Alastair
    • Look at Stephen's patches to updaterpms component
    • SL6
    • Order up extra memory for brendel (to check LCFG compiler performance)
    • Reboot schiff and split
    • Bring in openafs book
    • Investigate possible timing problem with FC/multipath

  • Chris
    • Officially start OpenAFS automation project
    • Reboot tobermory and cochin
    • Meet with Stephen to try test AFS install

  • Gordon

  • Stephen
    • SL6
    • Reboot trondra and rendall
    • Meet with Chris to try test AFS install
    • Mirror LCFG ISOs on DR server (should be able to rsync from lcfg.org)
    • PkgForge admin docs

-- AlastairScobie - 29 Mar 2011

Topic revision: r7 - 11 Apr 2011 - 14:33:09 - GordonReid
 
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback
This Wiki uses Cookies