Bug 987447 - App can't auto scale down when the HAproxy processes running on the head gear are terminated
Summary: App can't auto scale down when the HAproxy processes running on the head gear...
Keywords:
Status: CLOSED DEFERRED
Alias: None
Product: OpenShift Online
Classification: Red Hat
Component: Pod
Version: 2.x
Hardware: Unspecified
OS: Unspecified
medium
medium
Target Milestone: ---
: ---
Assignee: Rajat Chopra
QA Contact: libra bugs
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2013-07-23 11:58 UTC by Jianwei Hou
Modified: 2015-05-15 00:19 UTC (History)
0 users

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2013-07-23 19:55:31 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Jianwei Hou 2013-07-23 11:58:00 UTC
Description of problem:
Given a scalable application is scaled up to 2 gears, when the haproxy processes running on the head gear(first gear) is terminated(killed), the app can not successfully auto scale down after a certain time if there is no traffic to the app.

Version-Release number of selected component (if applicable):
On devenv_3540

How reproducible:
Always

Steps to Reproduce:
1. Create a scalable application, scale it up to 2 gears
2. ssh into the first gear(head gear) of the app, kill all haproxy processes
3. Do not access the app, wait for some time, then check that the app is auto scaled down

Actual results:
The app won't auto scale down

Expected results:
The app should auto scale down now that multiple load balancers are implemented.

Additional info:
The app will scale down if we don't terminate the haproxy processes on the head gear.

Comment 1 Jianwei Hou 2013-07-23 12:05:06 UTC
Also can not auto scale up when haproxy processes on head gear are killed.

Comment 2 Abhishek Gupta 2013-07-23 18:00:23 UTC
Multiple load balancers are not automatically implemented. The support is now there in the platform but the cartridge needs to be changed for that.

Based on the work being done in this release, we are now HA capable, just that HA is not provided out of the box.


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