Bug 66367 - How to define a connection timeout for health check only
Summary: How to define a connection timeout for health check only
Status: NEW
Alias: None
Product: Apache httpd-2
Classification: Unclassified
Component: mod_proxy_hcheck (show other bugs)
Version: 2.4.54
Hardware: All Linux
: P2 major (vote)
Target Milestone: ---
Assignee: Apache HTTPD Bugs Mailing List
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2022-11-29 10:34 UTC by jbrenaudin
Modified: 2022-11-29 13:25 UTC (History)
0 users


Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description jbrenaudin 2022-11-29 10:34:21 UTC 
As described in this request https://bz.apache.org/bugzilla/show_bug.cgi?id=60948, we want to define a timeout for health check only and if the health check request is too long then the hcfail counter is incremented without waiting the health check response. If hcfail counter equals hcfailed attribute so the member is disabled.

We saw this request on your bug tracker https://issues.redhat.com/browse/JBCS-448,  but when we tested with connectiontimeout, the timeout was never thrown and the member was still active.

Tested configuration :

ProxyHCTemplate memberTpl hcmethod=HEAD hcuri=/management/health hcinterval=3 hcpasses=1 hcfails=1
<Proxy balancer://cluster_prod>
 BalancerMember https://192.168.220.221:8080 connectiontimeout=2 hctemplate=memberTpl
 BalancerMember https://192.168.220.222:8080 connectiontimeout=2 hctemplate=memberTpl
</Proxy>

However we think the new attribute hconnectiontime proposed in https://bz.apache.org/bugzilla/show_bug.cgi?id=60948 will be more accurate instead of reusing the connectiontimeout attribute.
Comment 1 jbrenaudin 2022-11-29 13:25:13 UTC 
hconnectiontimeout instead of hconnectiontime