[
https://issues.apache.org/jira/browse/TS-1419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14622425#comment-14622425
]
Alan M. Carroll commented on TS-1419:
-------------------------------------
As far as I can tell this is as expected. If the proxy host is down, then it
will be retried after a while, specified by
proxy.config.http.down_server.cache_time
(https://docs.trafficserver.apache.org/en/latest/reference/configuration/records.config.en.html#proxy-config-http-down-server-cache-time).
This will generate the 502. If the down server is not retried at some point, I
don't see how it can ever be used again by ATS.
There is a complaint (and potential bug) that ATS does not failover internally
to another round robin address if the selected one is down. I'm investigating
that now.
> in strict round robin mode, some hostdb entry marked down(ip is down) will be
> used, then some request will be responsed 502
> ------------------------------------------------------------------------------------------------------------------------------
>
> Key: TS-1419
> URL: https://issues.apache.org/jira/browse/TS-1419
> Project: Traffic Server
> Issue Type: Improvement
> Components: DNS
> Affects Versions: 3.3.0, 3.2.0
> Environment: CONFIG proxy.config.hostdb.strict_round_robin INT 1
> Reporter: Bin Chen
> Assignee: Alan M. Carroll
> Labels: yahoo
> Fix For: 6.1.0
>
>
> in strict round robin mode, some hostdb entry marked down(ip is down) will be
> used, then some request will be responsed 502. Should we skip entry which be
> marked down? so, we can't always response by alive server(up entry).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
