[
https://issues.apache.org/jira/browse/TS-3724?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14606838#comment-14606838
]
ASF subversion and git services commented on TS-3724:
-----------------------------------------------------
Commit be68bd8f47f7ecde5403d9a63dbf81604d9bdf56 in trafficserver's branch
refs/heads/master from [~jacksontj]
[ https://git-wip-us.apache.org/repos/asf?p=trafficserver.git;h=be68bd8 ]
Make HostDBRoundRobin::select_best_http take last_failure time into
consideration for all RR types
In the current setup it only checks that status of the reals if you use
"default RR" (which is actually consistent hashing... but we'll let that
slide). This patch consolidates the alive() check into the HostDBInfo struct,
and then calls if from all 3 LB mechanisms. Since you can control if/when a
host is marked as down in ATS there is no reason to not check.
Issue: TS-3724
> ATS doesn't properly check last_failure time for most RR methods
> ----------------------------------------------------------------
>
> Key: TS-3724
> URL: https://issues.apache.org/jira/browse/TS-3724
> Project: Traffic Server
> Issue Type: Bug
> Reporter: Thomas Jackson
> Assignee: Thomas Jackson
>
> Right now strict and timed RR don't check the last_failure time of the real.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
