Bug 1251322 - op start timeout too short for httpd pacemaker resource
Summary: op start timeout too short for httpd pacemaker resource
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: openstack-foreman-installer
Version: Director
Hardware: Unspecified
OS: Unspecified
high
medium
Target Milestone: y2
: Installer
Assignee: Jason Guiditta
QA Contact: Shai Revivo
URL:
Whiteboard:
Depends On:
Blocks: track-osp6-osp7
TreeView+ depends on / blocked
 
Reported: 2015-08-07 01:23 UTC by Lars Kellogg-Stedman
Modified: 2016-05-20 15:09 UTC (History)
6 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2016-05-20 15:09:20 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Lars Kellogg-Stedman 2015-08-07 01:23:51 UTC
With a Staypuft deployed Pacemaker configuration, we have:

# pcs resource show horizon-clone
 Clone: horizon-clone
  Meta Attrs: interleave=true 
  Resource: horizon (class=systemd type=httpd)
   Attributes: start-delay=10s 
   Operations: monitor interval=60s (horizon-monitor-interval-60s)

That 10s start-delay can be too short given the pre-startup processes we run in /lib/systemd/system/httpd.service.d/openstack-dashboard.conf.  I think we should crank this up to 60s.

Comment 3 Lars Kellogg-Stedman 2015-08-07 04:36:50 UTC
So, s/start-delay/op start timeout/.  I was able to get reliable resoure startup by running:

  pcs resource op add horizon start interval=0 timeout=120

Comment 4 Mike Burns 2015-08-07 10:50:35 UTC
Moving to rhel-osp-installer components since this is staypuft.

Comment 5 Alan Pevec 2015-08-16 07:06:33 UTC
Adding Matthias to give guidance how long collectstatic + compress can take.

Comment 6 Matthias Runge 2015-08-17 05:35:23 UTC
I would even give it a little more, depending on IO speed. This compress step runs through all css/sass/less files, parsing and translating them.

On a slow machine, this still shouldn't take longer than 2 mins.

Comment 8 Jason Guiditta 2016-05-20 15:09:20 UTC
I don't believe any code fix will be made to adjust this at this point, reopen if it becomes a priority.  Otherwise, I think this is a manual step if needed for anyone on legacy setup.


Note You need to log in before you can comment on or make changes to this bug.