You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@aurora.apache.org by "Bill Farner (JIRA)" <ji...@apache.org> on 2014/09/09 18:29:28 UTC
[jira] [Commented] (AURORA-693) Aurora Client is pretty agressive
when server answers with anything but 200
[ https://issues.apache.org/jira/browse/AURORA-693?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14127183#comment-14127183 ]
Bill Farner commented on AURORA-693:
------------------------------------
There are two naive loops that are vulnerable to this:
- connecting
https://github.com/apache/incubator-aurora/blob/9f594f1de6bbf46c74863dd3fc4d2708b7a974f2/src/main/python/apache/aurora/client/api/scheduler_client.py#L241
- sending an RPC
https://github.com/apache/incubator-aurora/blob/9f594f1de6bbf46c74863dd3fc4d2708b7a974f2/src/main/python/apache/aurora/client/api/scheduler_client.py#L274
I agree that the retry behavior deserves some attention.
> Aurora Client is pretty agressive when server answers with anything but 200
> ---------------------------------------------------------------------------
>
> Key: AURORA-693
> URL: https://issues.apache.org/jira/browse/AURORA-693
> Project: Aurora
> Issue Type: Bug
> Components: Client
> Reporter: alexius ludeman
>
> When using aurora client and scheduler responds with non-200, the client goes in a pretty tight infinite loop of retrying with the following message:
> INFO] Starting new HTTP connection (1): $hostname
> There are a few ways to reproduce:
> -invalid dns name
> -a port that responds with "Connect Refused"
> -server responds with non-200 (specifically 401)
> Some ways forward:
> - the client should try X many times, and give up.
> - sleep between requests.
> - log why it is failing
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)