MPU Meeting Tuesday 27th August 2013

Inventory

Nothing happened. Alastair is planning to spend some more time on the project this week.

Sleep Enhancements

Stephen has manually culled any sleep data in the BuzzSaw DB which is older than our standard data retention limit of 120 days. This still needs to be automated.

The job of updating the daily sleep report so it does the necessary analysis needs to be added to the activity list.

Virtual DICE

Chris has now written a script for storing the necessary passwd and group data for users. The VM image is ready for testing, Alastair will test it. Chris will write some documentation for users. He will also write up his notes on how to build a new image in case we need to do so when he is not around.

LCFG Client Refactoring

Nothing happened this week.

Miscellaneous Development

server criticality
Stephen has been writing a script to power-off servers at a specified criticality level. This should work although it's hard to test without actually powering-off lots of machines. It still needs support for selecting a particular machine room. This will use a local cache of location information to avoid a dependency on querying the inventory DB.

LCFG portreserve support
Stephen has rolled out an update to the lcfg-ngeneric framework which adds support for configuring portreserve and running portrelease when a component is started. This should help daemons avoid having their ports stolen by the portmapper.

LCFG build tools and svn2cl
Stephen fixed bug#669 which meant that if you built a package from a branch and the changelog was generated from the subversion log you would get the log from trunk instead of the branch. There is some sort of side-effect problem which means that the generated changelog file does not contain the log entry for the most recently submitted changes, Stephen will investigate.

refreshpkgs
This has been converted to using the AFS::Command module. Yhr bucket creation script still needs converting but that is less urgent since we won't need it before SL7 and that could still be done manually if necessary. It might well be that the AFS::Command module needs some support adding for setting ACLs. Alastair discovered that the AFS::Command module needed an undocumented flag to force it to use the local timezone rather than just assuming GMT. At the moment brendel is still running openafs 1.4.14.1, Stephen will reboot it to get 1.6.5 between 08:00 and 09:00 on Wednesday. That gives us plenty of time to troubleshoot before the end of the week.

switchtoinfdb
Alastair has fixed the switchtoinfdb script so that the location data is up-to-date again.

kopenafs
Stephen has been investigating libkopenafs and the possibility of adding python and perl support.

KVM tools for users
Alastair looked at whether we could provide any simple KVM deployment/management tools for users of self-managed VMs. He concluded that for most systems are too heavyweight so for now we will just support the normal methods (e.g. virsh and virt-manager). We will make ISOs available for whichever platforms they need. We might also create the skeleton VMs with the required CPU/RAM/disk sizes. To begin with this isn't going to be widely publicised.

Operational

circle
Alastair has rebooted circle and there has been no repeat of the disk space issues.

computing.help slave
There is now a new computing.help slave server, named hjaelpe. This is a VM on hammersmith.

district
The problems on district were related to the removal of the SAN storage volume. There is now a new 50GB volume for testing purposes. Alastair will tidy up the remains of the two VMs which were lost during this process.

pkgsearch
Roger has removed all the duplicated configuration. He still needs to merge into the header all the SSL support which is currently only in the LCFG profile.

This Week

  • Alastair
    • Start Inventory project diary
    • Inventory project
      • Consider how to present design to CSOs - principally to ensure have covered every possible state and transition. Also to ensure not overly complicated to use
      • Consider technical talk
      • Submit bug/enh to App::Cmd author wrt option to die on unspecified options
      • Pester George about location API
    • Order a spare 600GB disk for waterloo.
    • Discuss NFS installroot problem with George - so why stopped working???Suspect that UDP was only just working and some change was sufficient just to tip us over the edge.
    • Ask George - what does the TXretransmit value mean for switch connections?
    • Consider how to make metropolitan usable by users
      • ISOs
      • minimal docs (mostly manual)
      • they'll use virt-manager, but not create machines or change config
    • circulate table of LCFG bugs
    • Discuss pkgsearch with Roger - wrt handover - Roger working on header.
    • Try out Chris's virtual DICE image
    • Remove old kvm guest config from district

  • Chris
    • Virtualised DICE image
      • Documentation - for users and MPU
    • Discuss autorebooting servers with RAT and implement

  • Stephen
    • Investigate mod_waklog behaviour when single des removed (using kerberos test cell)
    • Client refactoring project
      • complete writing more tests
    • Further investigation wrt AT / HP 8300 updaterpms issues
    • discuss release testing variants with Richardand then document.
    • Start looking at freenx
    • Stephen do brendel AFS upgrade
    • Flesh out LCFG config to git project (289)
    • Discuss Securing Web Servers (287) proposal with Craig

  • Carol
-- AlastairScobie - 27 Aug 2013
Topic revision: r8 - 03 Sep 2013 - 12:38:24 - 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