[jira] [Commented] (TC-317) ORT syncds bug - Parent holds up child running syncds if TO has multiple CDNs

["Hank Beatty (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TC-317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16126161#comment-16126161 ] 

Hank Beatty commented on TC-317:
--------------------------------

[~jpappa200] [~jvd] [~jeffmart] Is this still an issue? If so, does it need to be fixed in 2.1?

> ORT syncds bug - Parent holds up child running syncds if TO has multiple CDNs
> -----------------------------------------------------------------------------
>
>                 Key: TC-317
>                 URL: https://issues.apache.org/jira/browse/TC-317
>             Project: Traffic Control
>          Issue Type: Bug
>          Components: Traffic Ops, Traffic Ops ORT
>            Reporter: Nir Sopher
>              Labels: syncds
>
> Based on: https://github.com/Comcast/traffic_control/issues/1380
> By: jpappa200
> A parent from a different CDN can hold up a child from running syncds if Traffic Ops has multiple CDN's configured.
> https://<to_host>/update/ looks at all parents associated with the cache group and doesn't check which CDN it's associated with.
> Comment by jeffmart:
> Is this hold up indefinite?
> or
> does syncds complete on the child after all parents for all CDNs in that cache group are updated?
> I am assuming https://github.com/Comcast/traffic_control/pull/2 but want to make sure
> knutsel :
> so we have a issue somewhere to take regex_revalidate out of the ort/syncds distribution mechanism, and scp (or ansible dist) it directly on update; this would remove the "mids have to complete ort syncds before edges" requirement, and things should get a lot better when we do that...



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)