Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1562244

Summary: HAProxy keep-alive timeout exceeded by 5 seconds (docs)
Product: OpenShift Container Platform Reporter: Aleksandar Kostadinov <akostadi>
Component: DocumentationAssignee: Vikram Goyal <vigoyal>
Status: CLOSED CURRENTRELEASE QA Contact: Vikram Goyal <vigoyal>
Severity: medium Docs Contact: Vikram Goyal <vigoyal>
Priority: medium    
Version: 3.9.0CC: aos-bugs, bbennett, bmeng, bperkins, hongli, jokerman, mmccomas, zzhao
Target Milestone: ---   
Target Release: 3.10.0   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2018-07-09 00:29:30 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
ruby keep-alive test script
none
ruby keep-alive test script none

Description Aleksandar Kostadinov 2018-03-29 21:14:52 UTC
Created attachment 1414943 [details]
ruby keep-alive test script

Description of problem:
Setting ROUTER_SLOWLORIS_HTTP_KEEPALIVE to some number of seconds, e.g. 10s makes HAProxy close client connection after 15s. If the setting is 7s, then connection is closed after 12s of inactivity.

Version-Release number of selected component (if applicable):
3.9.14-1.git.0.4efa2ca.el7.x86_64

How reproducible:
always

Steps to Reproduce:
1. set variable in router dc ROUTER_SLOWLORIS_HTTP_KEEPALIVE=10s
2. wait router to restart
3. create some app in a user project
4. test keep-alive:
4.1. connect server
4.2. execute get command
4.3. wait

Actual results:
connection closes after 15 seconds

Expected results:
connection closes after 10 seconds

Additional info:
Please check attached ruby script to see how this is being tested.

Comment 1 Aleksandar Kostadinov 2018-03-29 21:19:00 UTC
Created attachment 1414944 [details]
ruby keep-alive test script

Comment 2 Jacob Tanenbaum 2018-04-17 15:40:35 UTC
Some of the timeouts are additive, in this case 

ROUTER_SLOWLORIS_HTTP_KEEPALIVE which adjusts "timeout http-keep-alive" triggers and then haproxy still waits on the "tcp-request inspect-delay" which we set to 5s 


I will update the docs mentioning that some of the timeouts interact so they may not be exactly as set

Comment 3 Aleksandar Kostadinov 2018-04-17 16:01:21 UTC
Thank you, we will adjust our tests according to docs when ready.

Comment 4 Jacob Tanenbaum 2018-04-27 16:55:09 UTC
https://github.com/openshift/openshift-docs/pull/8900

Comment 6 zhaozhanqi 2018-05-16 05:44:05 UTC
Checked the document PR, Verified this bug