Open main menu

CDOT Wiki β

Fedora ARM Meeting 2012 August 08

Revision as of 12:09, 8 August 2012 by Maxamaxim (talk | contribs) (Created page with 'Category:Fedora ARM Meeting =CDOT Fedora ARM Team Meeting 2012-08-08= = Updates = == Rebuild == * Not using full capacity - need to bring up at meeting this afternoon * Hit F…')
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

CDOT Fedora ARM Team Meeting 2012-08-08

Updates

Rebuild

  • Not using full capacity - need to bring up at meeting this afternoon
  • Hit F18 branch - need tags for ARM?

Pi

  • The 004 Image looks good - some reports of issues but can't recreate (?)
  • Swap creation seems reliable
  • Additional firstboot modules
    • Video configuration
    • Memory configuration
    • Keyboard config (language?)
  • Firstboot changes
    • Accelerator keys
  • Installer (NEED TAKER)
    • XZ compression - can we support this in the installer? pyliblzma
    • Mac support?
    • Windows priviledge support via manifest
    • Language support - parsing output of other commands
    • Command-line operation
  • Image shrink (Andrew)
    • 2G card
    • See Harold Hoyer's systemd boot optimization blog post

V6 Experiments

  • Build the Remix package set for armv6sf armv6hf (Jon + Chris)
    • Oscar to rebuild armv6hf
    • Felix to rebuild armv6sf

Lab

  • Need HP PC set up (Max)
  • Need cleanup for tour tomorrow morning (!)

Buildfarm

  • Icinga up and running!j
  • Jordan to do a demo/seminar soon

Packages

  • Look at the VC demos
  • Test gpio libraries
  • Look at the Adafruit image, see what libraries they are using
  • Scratch (Max + Chris)

maxam updates

  • Setup and Maintenance
    • WhiteyBoard in the Lab - Done (a million thanks to agreene!)
    • WhiteyBoard in other locations (all credits to agreene - certified WhiteyBoard expert)
    • Organizing the Pi+TV table in the Lab - ?
    • HP Touch based Computer as Fedora - Not done
  • No RTC Package
    • ctyler did an informal review
    • Did bug fixes and other fixes according to ctyler's suggestion
    • Ready to be deployed (with ctyler's approval) in the rpfr repo
    • Will attempt to get it in the main koji / arm koji - ?
  • RPFR Compose Script - rpfrcompose15.max(am)
    • Enhancements
      1. Removed older commented out commands - cleaner
      2. Reduced time to install and reinstall packages
      3. Removed rc.local based time fix
      4. Added nortc - as a yum localinstall
      5. Utilized image resizer script by ctyler
      6. Auto xz compression - slow conn. download happy size - faster burn rate
      7. Auto image availability as http://huttriver.proximity.on.ca/outbound/maxam.build.xz
    • Concerns
      1. Removed some commented out commands that were for debugging
      2. Outbound maxambuild image gets over written regardless of a good/bootable image compose when the compose script runs
      3. The xz command claims all the time recovered in the install process
      4. The image resizer script has mixed success
  • Backup (hoping everyone read the email from yesterday - if not, please read it)
    • Now: Stick to the current version of blaze-wildfire
    • Test1: Newer version of blaze-wildfire which envokes rsync in a different way
      • Recommandations
        1. Test the newer blaze-wildfire
        2. Go through a test phase (2 or 3 days)
    • Test2: Longterm (and necessary) solution is to add lvm and lv snapshot to get a better (less volatile) copy of the repo
      • Concerns
        1. Minor possibility of a performance hit
        2. Cost $1850+ CDN ($1300 for the 600G SSD and $170x2 for the 3T drives)
        3. Setup complications - remove raid, add drives, new raid, do lvm and so on - reasonable amonut of smaller steps in between that could go wrong
        4. Requires chile to be down - buildfarm will be down
      • Recommandations (after Test1)
        1. Test the lv snapshot scenario in scotland (ctyler graciously agreed to supervise ) - less chance of maxam destroying scotland :)
        2. Go through another test phase
        3. Schedule downtime -- let the channel & list know ahead of time
        4. Go thru hardware upgrade in chile and then adding the lv and snapshot process
        5. At anypoint if anything goes wrong - scream like a little child and go seek help from ctyler