aurora-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dmitriy Shirchenko <cald...@gmail.com>
Subject Re: Review Request 47440: RFC for WIP: adding webhook code that can POST events to an endpoint
Date Wed, 18 May 2016 21:12:33 GMT


> On May 17, 2016, 2:51 a.m., George Sirois wrote:
> > src/main/java/org/apache/aurora/scheduler/events/Webhook.java, line 86
> > <https://reviews.apache.org/r/47440/diff/1/?file=1385132#file1385132line86>
> >
> >     Should this be dispatched on another thread or added to a queue for future delivery?
AFAICT the distribution of events is synchronous, so it seems like operations that might block
for any significant amount of time could cause issues.
> 
> Dmitriy Shirchenko wrote:
>     Yea, I'm not sure. Do you have examples of best practices on how to dispatch to another
thread?
> 
> Dmitriy Shirchenko wrote:
>     I looked at the logs, and looks like each event is sent via own process judging by
logs. 
>     
>     Do you think this is correct or I'm off base? 
>     ```
>     I0517 06:28:28.250 [AsyncProcessor-6, Webhook:57] Sending message {"task":{"cachedHashCode":0,"assignedTask":{"cachedHashCode":0,"taskId":"vagrant-test-http_e
>     I0517 06:28:28.254 [AsyncProcessor-6, Webhook:77] Disconnecting 
>     I0517 06:28:28.256 [AsyncProcessor-6, Webhook:80] Done with call 
>     <snip>
>     I0517 06:28:28.808 [AsyncProcessor-4, Webhook:57] Sending message {"task":{"cachedHashCode":0,"assignedTask":{"cachedHashCode":0,"taskId":"vagrant-test-http_e
>     xample-0-81ba7ce9-1900-4177-98af-847ae3a90fb7","slaveId":"8639d42c-2c68-40d9-9131-43e597457de5-S0","slaveHost":"192.168.33.7","task":{"cachedHashCode":-528474
>     <snip>
>     I0517 06:28:28.810 [AsyncProcessor-4, Webhook:77] Disconnecting 
>     I0517 06:28:28.810 [AsyncProcessor-4, Webhook:80] Done with call 
>     ```
> 
> Maxim Khutornenko wrote:
>     We use AsyncEventBus with the thread pool size of 8: https://github.com/apache/aurora/blob/b24619b28c4dbb35188871bacd0091a9e01218e3/src/main/java/org/apache/aurora/scheduler/async/AsyncModule.java#L63
>     
>     It should be ok to call on an event bus thread as long as connection timeouts are
set aggresively (see Stephan's suggestion below).

Sweet. Can I close this out as Fixed?


- Dmitriy


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/47440/#review133482
-----------------------------------------------------------


On May 18, 2016, 9:12 p.m., Dmitriy Shirchenko wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/47440/
> -----------------------------------------------------------
> 
> (Updated May 18, 2016, 9:12 p.m.)
> 
> 
> Review request for Aurora.
> 
> 
> Bugs: AURORA-1683
>     https://issues.apache.org/jira/browse/AURORA-1683
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> Looking for some feedback whether I'm on the correct path in adding a webhook. All comments
are welcome!
> 
> 
> Diffs
> -----
> 
>   docs/reference/scheduler-configuration.md f7d676d0ed6bc536f4341dbb9365cf50e8607efb

>   examples/vagrant/upstart/aurora-scheduler.conf 3d9e706de564df5e24cb34265bebc0db1cad11a0

>   src/main/java/org/apache/aurora/scheduler/app/SchedulerMain.java 9ebfe230836e88a97bc60092373f72f176a8f6f2

>   src/main/java/org/apache/aurora/scheduler/events/PubsubEvent.java 2a4c0665e48d30e0655de00bd7f6f9b49f01eafc

>   src/main/java/org/apache/aurora/scheduler/events/Webhook.java PRE-CREATION 
>   src/main/java/org/apache/aurora/scheduler/events/WebhookInfo.java PRE-CREATION 
>   src/main/java/org/apache/aurora/scheduler/events/WebhookModule.java PRE-CREATION 
>   src/main/resources/org/apache/aurora/scheduler/webhook.json PRE-CREATION 
>   src/test/java/org/apache/aurora/scheduler/events/WebhookTest.java PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/47440/diff/
> 
> 
> Testing
> -------
> 
> Need to fix tests.
> 
> 
> Thanks,
> 
> Dmitriy Shirchenko
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message