Bug 1260844 - Horizon's "Resource Usage" tab hangs when ceilometer has a lot of data
Summary: Horizon's "Resource Usage" tab hangs when ceilometer has a lot of data
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: python-django-horizon
Version: 7.0 (Kilo)
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
: 8.0 (Liberty)
Assignee: Matthias Runge
QA Contact: Ido Ovadia
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2015-09-08 03:29 UTC by August Simonelli
Modified: 2015-09-08 05:31 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2015-09-08 05:31:05 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description August Simonelli 2015-09-08 03:29:51 UTC
Description of problem:
When you click on "Resource Usage" the browser eventually times out (after 60s) with an error and doesn't show anything.

Version-Release number of selected component (if applicable):


How reproducible:
Have a lot of samples in ceilometer, then go to Horizon and click the Resource Usage tab. See https://bugzilla.redhat.com/show_bug.cgi?id=1260842

Steps to Reproduce:
1. As above.
2.
3.

Actual results:
Browser times out

Expected results:
Returns ceilometer data.

Additional info:
It seems this is cause ceilometer is slow to return data. See https://bugzilla.redhat.com/show_bug.cgi?id=1260842

Comment 3 August Simonelli 2015-09-08 03:36:42 UTC
I hit this in conjunction with https://bugzilla.redhat.com/show_bug.cgi?id=1260842

I figured if i changed all my endpoints to bypass haproxy, ie:

[root@overcloud-controller-0 ~ (overcloud-site2)] $ grep -A 5 ceilometer /etc/haproxy/haproxy.cfg
listen ceilometer
  timeout  server 2m
  bind 10.0.1.20:8777
  bind 10.0.14.20:8777
  server overcloud-controller-0 10.0.1.23:8777 check fall 5 inter 2000 rise 2

and

[root@overcloud-controller-0 ~ (overcloud-site2)] $ openstack endpoint show 249ecf4fd1fa4f5db8f8e81dc566ed72
+--------------+----------------------------------+
| Field        | Value                            |
+--------------+----------------------------------+
| adminurl     | http://10.0.1.23:8777            |
| enabled      | True                             |
| id           | 249ecf4fd1fa4f5db8f8e81dc566ed72 |
| internalurl  | http://10.0.1.23:8777            |
| publicurl    | http://10.0.1.23:8777            |
| region       | regionOne                        |
| service_id   | 2180765d499d432e86c7d89448174cc5 |
| service_name | ceilometer                       |
| service_type | metering                         |
+--------------+----------------------------------+

then perhaps horizon would be happy (if only for a test).

No luck. Horizon still dies after 60 seconds with "The connection was reset"
If i ask for /admin/metering/anything i do get a 404.
When asking for /admin/metering/ i see no access requests in the logs. Which makes sense combined with the client hanging waiting for ceilometer to return.

Also, ceilometer sample-list is slow:

[root@overcloud-controller-0 ~ (overcloud-site2)] $ time ceilometer sample-list| wc -l
135519

real    2m54.349s
user    0m59.231s
sys     0m1.548s

Comment 4 August Simonelli 2015-09-08 05:31:05 UTC
ok, so i've changed the haproxy timeout for the vip i use to connect to horizon and that fixed it. this is not a bug. i'll close it.


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