MPU Meeting Tuesday 6th March 2012

LCFG Server Refactoring

Trondra has been upgraded to be the same as mousa but is not displaying similar rsync errors. Mousa is in the tower so it may be a networking problem. There are three other test slaves; vole, benaulim and sernabatim. Only the last shows the occasional rsync error. Suggestion is to run circlevm12 as test profile server since circle is in the tower

Simple KVM Service

Waiting for rvirsh script from Iain. More documentation has been added.

SL6 Server Upgrades

Chris tried to make 'tasmin' be an rfe server but without much luck. George and Ian to investigate a certificate error. We should chase up other units progress on porting defaults packages, there are 13 still to go and 5 in the devel bucket when they should be in the LCFG bucket. External sites can't see the devel bucket. The ordershost side of tobermory is done. Alastair is planning a schema change. The rfe configuration is important as it holds the master inventory files, which are synced to the database. The web stuf is less crucial. The manual steps for bootstapping the ordershost are in dice/options/ordershost.h (at the bottom). Chris and Alastair to liaise. Other SL6 upgrades are blocked by the need for AFS (1.6.1 rc 4 is close to what we need ) Stephen is now on the AFS release-team mailing list.

Miscellaneous Development

  • lcfg-sleep : Chris has picked up some tips from James in IS to do with sleep/wake/wol, but hasn't had a chance to try them yet.

  • New HPs : The 8200 was previously was previously tested as a selectPC. Intended as lab-machine replacement.
It needs tested with sleep, as Chris thinks it doesn't respond to mouse/keyboard only power-switch to wake up. Going back to the the standard small-form factor not towers, and no ultra-small form factor.

Operational

  • lochranza : replaced memory (borrowed from ratz) - needs upgraded soon.

  • Flood drill : see Deficiencies section of the Unit report. Discussion pointed out the need for 'procedures' to be documented as part of the SL6 upgrades project. We should add a column to the ServersList with a link to a 'restoration procedure' document for each. It was noted PXE installs for SL5 were broken. These don't get tested as part of the weekly tests. Introduce versioning to tftpboot kernel files.

This Week

  • Alastair
    • Nag Iain re rvirsh script
    • Tidy up dice-orders header for SL6
    • arrange figgy to go to KB
    • Improve dice-orders reinstall docs
    • Add doc on reinstalling a KVM and its guests
    • Cron job to run on KVM servers which emails mp-unit with which guests it hosts
    • Add support for SL5 on KVM servers
    • For operational : record in LCFG profile what resources are being used - eg SAN, VM host etc
    • Review project list
    • Start thinking about inventory project

  • Chris
    • Configure one of circle's VMs to be an LCFG slave to rule out AT network issues wrt mousa rsync errors
    • Kick off Server Hardware interaction project
    • Add action on each service with data in SL6ServerUpgradeList to update relevant section in MPUBackupSummary
    • Sort out DICE_NO_NAGIOS in MPU profiles
    • Review project list
    • SL6 defaults list
    • PD - local url shortener implementation (for not-a-service)

  • Stephen
    • Write UKUUG talk
    • Theon
    • Add SL6 epel to sauce
    • Review project list
    • PXE installroot - check documentation about moving link when building pxe root
    • Consider PD - concrete task

-- AlastairScobie - 06 Mar 2012

Topic revision: r11 - 19 Mar 2012 - 16:35:11 - ChrisCooke
 
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