MPU Meeting Tuesday 4th August 2015

LCFG Client Refactoring

Nothing happened.


Alastair has written up the meeting he had with Graham and Chris.

Alastair will handle the conduits the way Graham suggests - from Theon to CSV, uploading to Tartarus with triggers. It will be necessary to formally specify exactly what information to pull from Theon. It will also be necessary for Theon to handle the "space" concept properly. Once that's done it'll be possible to get the rest of the prototype up and running again.

The prototype code now includes rack location (meaning which rack the server is in - useful for servers). We talked about whether rack locations should be regarded as rooms or as another field, and concluded that they should be another field. We further concluded that the field should be generalised to "sublocation" so that it can also cover e.g. which desk in a large office a desktop might be at.

Alastair has also been considering the MAC discovery tool (which might as well grab other relevant info too while it's at it). We could have a PXE option, with the installroot, which registers the machine in the inventory. Alastair will attempt this in a generic way. He'll make sure that the PXE option is hidden on MDP machines.

Authentication is another consideration. It needs to work on MDP as well as DICE. Stephen suggested a Cosign-authenticated web page. If the schema changed, we'd simply need to change the web backend. Alastair will check this idea with the CSOs as he doesn't want to burden them with awkward tasks.

Miscellaneous Development

Stephen moved the trigger script for rebuilding the initramfs file from the kernel component into the main kernel package. Hopefully this will sort the problem of the locale and keyboard map being wrong at install time. Stephen will change the resources for the kernel component to add dependencies

Stephen is going to update the SL7 PXE installroot to be the latest one with the Network Manager fix.

For 7.1 we need to force a reboot on minor release upgrade. lcfg-reltool minor lcfg-release should guarantee a reboot after the install has finished.

Stephen fixed the installbase to have the correct SSL certificate for Edinburgh University. He's introduced ed/options/installbase.h to do this.

Alastair noticed that when building the installbase, SSL certificate installation fails. He'll create a bug.

Alastair has altered kvmtool to support the inifile configuration format. Chris is going to produce new configuration to match. The latest kvmtool will also be happy if no default bridge is defined, so we'll be able to go ahead and remove br0 from the KVM server configurations.

Alastair has documented how to find packages. It's now in the Packages section of the MPU documentation.


Chris cleared up the NX server as it was full of old student processes left over from the spring. We should try to make more of an effort to make the NX servers more self-tidying and self-maintaining.

Alastair rebooted hare, wildcat and northern and is scheduling a reboot of oyster.

Alastair has ordered extra memory for KVM servers oyster and hammersmith to take them up to the 128GB memory enjoyed by most of the other KVM servers. Stephen has the memory and will install it at 6.6 upgrade time.

Stephen spent some time working on Phil Wadler's machine since its monitor setup is somewhat fragile at reboot time. He's looking at ways of making it less so. New monitors may help; upgrading to the latest NVidia driver helped too.

Stephen drew our attention to OpenAFS 1.6.13, a security release. It's on DICE but as yet still optional. The security fix contained in 1.6.13 is to remove most regexp support, but some sites depend on this for backup - do we?

Chris and Stephen will re-rack the new KVM server azul in an FC rack in case we ever want to use FC with it. While it's down Alastair wants to measure its power consumption at idle, since it's a new model to us.

SL 6.7 is imminent. Producing the upgrade should only take a few days this time. However, we'll first need to get rid of 6.5 entirely because we simply don't have the resources to support three SL6 minor releases at one time.

We need to flesh out the SL7 server project.

Stephen has taken a look at the xinetd component. It needs rewriting into Perl. For now he'll patch it to get it working on SL7.

This Week

  • Alastair
    • Inventory project
      • continue working through TartarusWorkFlow
      • finish deploying new order file processing code
      • finish document new parser and API
      • consider what next can be integrated into existing system, if anything
      • write up meeting with Graham re conduits Awaiting Chris and Graham's comments
      • Consider sublocation, macaddr discovery tool
    • @home - look at using rsync from site.pkgs instead of mirroring from upstream
    • Remove default pool if ops meeting agrees
    • Schedule firmware upgrade for DS3254
    • Take creation of mdp-installbase profiles to LCFG deployer's meeting (bug #878) (On July meeting agenda)
    • Add instructions on adding extra disks to an LVM/KVM pool - to KVM host guide
    • Talk to Phil Wadler about machine upgrade - ask for a list of software he requires built
    • Experiment with different window managers under VNC (making the assumption that performance under NX will be similar)
    • Schedule reboot of oyster with new memory install - Expand suspend space- needs a reboot to pick up /dev/sda4 - move /var/libvirt before KVM clients come up -await new memory Check using eth0 and eth2
    • Think of a use for 'atom'
    • Ship new version of kvmtool
    • Understand how NetworkManager works wrt init scripts
    • Look at RT 72904 We need to maintain a h/w page per model + should look at distributing BIOS settings via USB sticks
    • Flesh out Base SL7 server project

  • Chris
    • Inventory project
      • continue working through TartarusWorkFlow
      • code review parser
      • Comment on Graham/Alastair/Chris meeting notes
    • Finish work on SL 7.1
      • test installer after stable release of 05/08 (to check network bug gone)
    • pkgsearch for SL7
    • Expand suspend space on waterloo
    • Coordinate move of district to junk rack and harvest disks
    • Look at inifile component configuration of kvmtool
    • Start looking at SL7 Virtual DICE
    • Reboot piccadilly and bruegel
    • Add improved NX culling script to todo-list
    • Move azul
    • Flesh out Base SL7 server project

  • Stephen
    • LCFG client refactor stage 1
      • schedule debrief meeting
    • LCFG client refactor stage 2
      • finish off C code
    • Think about PD - Interested in ZeroMQ
    • Finish off window manager selector code
    • Reboot brendel and schiff
    • Add dependency for initramfs rebuild on kernel, dracut and systemd (ASAP)
    • Update PXE installroot (for new lcfg-network)
    • Create a bug wrt eucs-certs install failing (in installroot)
    • Move azul
    • Flesh out Base SL7 server project

-- AlastairScobie - 04 Aug 2015

Topic revision: r12 - 23 Sep 2019 - 13:33:37 - AlastairScobie
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