yahoo-eng-team team mailing list archive

[Radosław Piliszek <1871138@xxxxxxxxxxxxxxxxxx>]

Affects kolla - flaky images.

** Also affects: kolla
   Importance: Undecided
       Status: New

** Also affects: kolla/stein
   Importance: Undecided
       Status: New

** Also affects: kolla/train
   Importance: Undecided
       Status: New

** Changed in: kolla/stein
       Status: New => Triaged

** Changed in: kolla
       Status: New => Invalid

** Changed in: kolla/train
       Status: New => Triaged

** Changed in: kolla
   Importance: Undecided => High

** Changed in: kolla/stein
   Importance: Undecided => High

** Changed in: kolla/train
   Importance: Undecided => High

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Dashboard (Horizon).
https://bugs.launchpad.net/bugs/1871138

Title:
  Horizon taking a long time to compress static assets

Status in OpenStack Dashboard (Horizon):
  New
Status in kolla:
  Invalid
Status in kolla stein series:
  Triaged
Status in kolla train series:
  Triaged
Status in kolla-ansible:
  Invalid
Status in kolla-ansible stein series:
  Triaged
Status in kolla-ansible train series:
  Triaged

Bug description:
  Currently in Kolla Ansible CI we are hitting an issue where accessing
  the dashboard times out. We have a timeout of 100s. The container is
  up and running, with no errors. Checking the ps logs, we can see the
  CPU is 100% performing static asset compression:

  root     31399 31333 31399 99.1  3.7 336828 302972
  /var/lib/kolla/venv/bin/python /var/lib/kolla/venv/bin/manage.py
  compress --force

  I did some binary-chopping of Python packages between the Stein and
  Train centos-source-horizon images. The culprit seems to be PyScss
  1.3.7. In the train image it is 1.3.4. On my laptop, this ~doubles the
  time required for static asset compression from 35s to 1m10s.

To manage notifications about this bug go to:
https://bugs.launchpad.net/horizon/+bug/1871138/+subscriptions