Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

  • Re-triggering Tests

    • All re-triggers (if you expect them to vote to Gerrit/Github must be done from the component flow jenkins  job)
    • Watch the #smoke-test channel for failures, this has a link to the component flow and change for each failure
    • Only need to re-trigger on main dev branches
      • master
      • rdk-next
      • stable2
      • dev_sprint_23_1
        • note stuff may fail here, so we just re-trigger to make sure of status, if everything is failing ping aamp team
      • If something gone wrong with jobs/lab Alan should investigate
  • Current Known Failures

    • Flashing
      • sometimes we lose telnet connection to iolan during flashing phase of flows, just need to re-trigger
    • Turris
      • sometimes the serial port is unreliable, just need to re-trigger
    • RPI-B
      • Sometimes flashing fails with ccspwifiagent, hostapd
        • Re-trigger if it fails again, it needs investigation
          • May need a manual verification build or Keerthana help
      • 5Ghz wifi failures these seem to occur when rdkb is busy, either override the Verified vote or re-trigger
      • WPA supplicant restart failures on our wifi clients  – have to re-trigger
      • Web UI Issues probably need input from Keerthana
    • RPI-V
      • Seeing AAMP-CI tests failing on mediaclient, either re-trigger or override the verified vote
      • There is an occasional bug where sometimes everything passes on re-runs within the test but Verified vote still -1, I haven’t had time to track this down.
    • BRCM refboard LGI refapp
      • Occassionally the home screen check fails (video not playing on boot)
        • Just re-run or override the Platform Verified label
    • Amlogic/Realtek
      • Occasionally but rarely these fail, normally work on re-trigger
      • If it’s consistently failing need to look at state of device as it may have become corrupt during flashing, or if a build was flashed where IP was not getting assigned it will need manual recovery from me
      • if in a bad state just reserve the device from LOCKABLE_RESOURCES on Jenkins as we have 3 of each in the racks and I will sort when back
  • Device in a bad state

  • Nightly CI

 

  • Disable Sanity on a Platform

    • If you want to disable sanity for a platform you need to turn off UPLOAD_IMAGES in the build jobs (DSL)
    • We normally don’t do this … last resort…

...