[
https://issues.apache.org/jira/browse/TS-570?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Leif Hedstrom updated TS-570:
-----------------------------
Actually, I think this subject is slightly misleading, the code just deal with
DNS threads > 1. That was actually intentional, but this patch is better than
what we used to have (although, not sure it ever makes sense to have more than
one DNS thread, that would need to be a hell of a lot of DNS requests). The
stuff around initialization was wrong though, so good catch on that.
In any case, the patch looks good, +1.
> SplitDNS schedules on wrong thread when "dns_threads > 0"
> ---------------------------------------------------------
>
> Key: TS-570
> URL: https://issues.apache.org/jira/browse/TS-570
> Project: Traffic Server
> Issue Type: Bug
> Components: DNS
> Affects Versions: 2.1.5
> Environment: 2.1.5 trunk
> Reporter: qianshi
> Assignee: Leif Hedstrom
> Priority: Minor
> Fix For: 2.1.5
>
> Attachments: TS-570.patch
>
>
> In main() function, SplitDNSConfig::startup() is called before
> dnsProcessor.start() like 1.
> In SplitDNSConfig::startup(), reconfigure() schedule threads like 2.
> In dnsProcessor.start(), create threads for dns and split dns like 3.
> So when split dns is enable and dns threads more than 0, split dns uses no
> the dns threads which created to do so but 0 thread, that may be not the
> initial purpose of creating the dedicated threads for dns and split dns.
> 1.
> SplitDNSConfig::startup();
> dnsProcessor.start();
> 2.
> EThread *thread = eventProcessor.eventthread[ET_DNS][0];
> thread->schedule_imm(dnsH, ET_DNS);
> 3.
> ET_DNS = eventProcessor.spawn_event_threads(dns_threads);
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
