MPU Meeting Tuesday 29th September 2015

Inventory

Alastair has been looking at using the jsonb support in PostgreSQL 9.4, sadly our version of the DBIx::Class module is too old and it does not work nicely. Instead we will just store the client report data in a TEXT field for now.

The client reporting is now done via an authenticated submission system. The new client has been deployed to develop machines and we are getting about 90 daily reports. The error handling for probe modules needs to be improved, we should decide where failure reports should be sent. The client should support using a configuration file so we can easily control which probe modules are run. Although functional the clientreport tools should still be considered to be at the prototype stage.

Alastair will now return to looking at the orders loading code.

LCFG Client Refactoring

The XS code for the LCFG::PkgSpec module has been reworked to use XS::Object::Magic so that the class is now fully sub-classable. The LCFG::Profile::PackageSpec module has also been reworked so that it is a sub-class of LCFG::PkgSpec with the Context and Derivation roles applied. Nearly all required functionality for processing package specifications has been pushed into C with XS wrappers so that it is reusable. The handling of package lists was also improved using a new LCFGPackageList struct to hide the details. All the package APIs have been documented.

SL7 Server Base

Networking
Stephen has confirmed that setting NM_CONTROLLED=no in the network interface config files allows us to use the old networking scripts to setup bonding, bridging and VLANs. We still need to convert over to using networkmanager at some point in the future but this will allow us to start getting servers upgraded.

IPMI
Chris has checked this with Ian D and it all looks fine.

Partition layout
This seems to be fine for both normal and small servers.

Nagios
Stephen has checked that the basic active nagios checks (e.g. openssh) work. He has also got the necessary packages built so that the passive network bonding monitoring script works correctly, it has been fully tested by disabling ports.

hwmon and toohot
Chris will look at hwmon and toohot monitoring.

localhome
There is currently a conflict between localhome and autofs related to the management of a /home symlink. This will have to be sorted out with Craig before we can upgrade many of our MPU servers.

Miscellaneous Development

Virtual DICE
Chris has a new SL7-based virtual DICE system working, it is just waiting on the final few RAT packages.

lightdm
Stephen and Graham had a hackathon to fix the problems with session handling in lightdm on SL7. The results are in a new dice-desktop package. The user can now make a permanent choice of which desktop they want using the switchdesk tool.

bashdefenv problems with gnome
Alastair is having some weird problems with bashdefenv in Gnome3 which no-one else is seeing.

pkgsearch
Chris is looking at assembling yum configurations for our platforms which the pkgsearch web server can use to do searches.

disk encryption
Alastair has modified hackparts and the fstab component to add support for disk encryption with /tmp and swap. Will look at deploying once he is back from holiday. Easy to enable for new machines but how should we approach applying this to existing machines? See circlevm12 for details, needs a special PXE installroot.

network-online hang
The boot-time hang with network-online has been fixed by re-enabling networkmanager and instead just setting NM_CONTROLLED=no in the network interface configuration files.

Operational

bugzilla
Chris has a test version of the bugzilla upgrade available at testbugs.lcfg.org, we should all check it out.

RT tickets
It looks like we're accumulating tickets again which could be easily closed.

SL6.5
We still need to remove the final traces of the DICE_STICK_WITH_SL65 macro from the LCFG source profiles.

This week

  • Alastair
    • Inventory project
      • continue working through InvProjectWorkFlow
      • finish deploying new order file processing code - just a matter of removing an override in steen's profile
      • consider what next can be integrated into existing system, if anything
      • Consider macaddr discovery tool (use same code as used for clientreport)
      • Polish clientreport in light of results returned from clients
      • Revisit order -> inventory sync code
    • @home - look at using rsync from site.pkgs instead of mirroring from upstream
    • Remove default pool if ops meeting agrees
    • Experiment with different window managers under VNC (making the assumption that performance under NX will be similar)
    • Think of a use for 'atom'
    • Understand how NetworkManager works wrt init scripts
    • Deploy encrypted /tmp and swap
      • Develop a script to modify existing machines
        • possibly just a matter of working out the PARTUUID of the swap and /tmp partitions and then editing the fstab and crypttab template files and scheduling a reboot?
      • Test <live/fstab_test.h> with an install
    • Look at RT tickets to close
    • SL7 base server
      • localhome - discuss with Craig
      • Look at metropolitan with SL7
    • investigate bash defenv problem -> RT ticket -> RATTurns out to be gnome-terminal started from Right-Click isn't being started in a manner that inherits the login environment. Starting gnome-terminal from Applications menu works fine
    • Look at testbugs
    • Produce list of displaced servers

  • Chris
    • Inventory project
      • continue working through InvProjectWorkFlow
      • Look at clientreport modules for replacing firmwarereport
    • pkgsearch for SL7
      • reimplement as a yum web front end (yum search for keyword produce an html file of links to cgi to do yum info)
      • Need support multiple platforms
    • Continue work on Virtual DICE for SL7
    • Continue work on bugzilla upgrade
    • Liaise with George over iDRAC documentation
    • SL7 -
      • hwmon
      • toohot
      • Look at DL180 (or alternative model)
    • RT tickets close
    • Start work on the spending plan for 2015/2018
      • identify any displaced (from other units) servers that we could use

  • Stephen
    • LCFG client refactor stage 1
      • schedule debrief meeting
    • LCFG client refactor stage 2
      • deploy package libraries and new whererpms
      • test under MacOS
      • work on resources processing
    • Think about PD - Interested in ZeroMQ
    • Write up how WM switchdesk mechanism works
    • Liaise with Roger over borrowing an R620 for SL7 testing
    • RT tickets close
    • Look at creating a new develop PXE root with new version of fstab component
    • Investigate kernel component pipe moan by using shell commands instead of RPM module
    • Look at testbugs
    • Sort out DICE_STICK_WITH_SL65

-- AlastairScobie - 29 Sep 2015

Edit | Attach | Print version | History: r13 | r10 < r9 < r8 < r7 | Backlinks | Raw View | Raw edit | More topic actions...
Topic revision: r8 - 12 Oct 2015 - 12:08:45 - 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