You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@aurora.apache.org by Bill Farner <wf...@apache.org> on 2014/09/20 05:06:03 UTC
Review Request 25872: Instrument task scheduling code to give
visibility into failed searches and backoffs.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25872/
-----------------------------------------------------------
Review request for Aurora, Kevin Sweeney and Maxim Khutornenko.
Repository: aurora
Description
-------
I'd like to use this data to investigate potential changes to the backoff strategy used in TaskGroups. I have a hunch that tasks are usually either scheduled on the first try, or after the 10th. If that is true, we can make difficult-to-schedule tasks much less expensive.
Diffs
-----
src/main/java/org/apache/aurora/scheduler/async/TaskGroups.java f64497fcf082defd4a6a0935febce8ed90c7e078
src/main/java/org/apache/aurora/scheduler/async/TaskScheduler.java 882cdfd9b79b262befb81437cbd9a31a6bc1e40f
Diff: https://reviews.apache.org/r/25872/diff/
Testing
-------
./gradlew build -Pq
Thanks,
Bill Farner
Re: Review Request 25872: Instrument task scheduling code to give
visibility into failed searches and backoffs.
Posted by Bill Farner <wf...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25872/
-----------------------------------------------------------
(Updated Sept. 22, 2014, 6:42 p.m.)
Review request for Aurora, Kevin Sweeney and Maxim Khutornenko.
Repository: aurora
Description
-------
I'd like to use this data to investigate potential changes to the backoff strategy used in TaskGroups. I have a hunch that tasks are usually either scheduled on the first try, or after the 10th. If that is true, we can make difficult-to-schedule tasks much less expensive.
Diffs (updated)
-----
src/main/java/org/apache/aurora/scheduler/async/TaskGroups.java f64497fcf082defd4a6a0935febce8ed90c7e078
src/main/java/org/apache/aurora/scheduler/async/TaskScheduler.java 882cdfd9b79b262befb81437cbd9a31a6bc1e40f
Diff: https://reviews.apache.org/r/25872/diff/
Testing
-------
./gradlew build -Pq
Thanks,
Bill Farner
Re: Review Request 25872: Instrument task scheduling code to give
visibility into failed searches and backoffs.
Posted by Bill Farner <wf...@apache.org>.
> On Sept. 22, 2014, 5:34 p.m., Kevin Sweeney wrote:
> > src/main/java/org/apache/aurora/scheduler/async/TaskScheduler.java, line 180
> > <https://reviews.apache.org/r/25872/diff/1/?file=698588#file698588line180>
> >
> > typo in "attemptsFired"
Thanks, fixed.
- Bill
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25872/#review54145
-----------------------------------------------------------
On Sept. 20, 2014, 3:06 a.m., Bill Farner wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25872/
> -----------------------------------------------------------
>
> (Updated Sept. 20, 2014, 3:06 a.m.)
>
>
> Review request for Aurora, Kevin Sweeney and Maxim Khutornenko.
>
>
> Repository: aurora
>
>
> Description
> -------
>
> I'd like to use this data to investigate potential changes to the backoff strategy used in TaskGroups. I have a hunch that tasks are usually either scheduled on the first try, or after the 10th. If that is true, we can make difficult-to-schedule tasks much less expensive.
>
>
> Diffs
> -----
>
> src/main/java/org/apache/aurora/scheduler/async/TaskGroups.java f64497fcf082defd4a6a0935febce8ed90c7e078
> src/main/java/org/apache/aurora/scheduler/async/TaskScheduler.java 882cdfd9b79b262befb81437cbd9a31a6bc1e40f
>
> Diff: https://reviews.apache.org/r/25872/diff/
>
>
> Testing
> -------
>
> ./gradlew build -Pq
>
>
> Thanks,
>
> Bill Farner
>
>
Re: Review Request 25872: Instrument task scheduling code to give
visibility into failed searches and backoffs.
Posted by Kevin Sweeney <ke...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25872/#review54145
-----------------------------------------------------------
Ship it!
src/main/java/org/apache/aurora/scheduler/async/TaskScheduler.java
<https://reviews.apache.org/r/25872/#comment94139>
typo in "attemptsFired"
- Kevin Sweeney
On Sept. 19, 2014, 8:06 p.m., Bill Farner wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25872/
> -----------------------------------------------------------
>
> (Updated Sept. 19, 2014, 8:06 p.m.)
>
>
> Review request for Aurora, Kevin Sweeney and Maxim Khutornenko.
>
>
> Repository: aurora
>
>
> Description
> -------
>
> I'd like to use this data to investigate potential changes to the backoff strategy used in TaskGroups. I have a hunch that tasks are usually either scheduled on the first try, or after the 10th. If that is true, we can make difficult-to-schedule tasks much less expensive.
>
>
> Diffs
> -----
>
> src/main/java/org/apache/aurora/scheduler/async/TaskGroups.java f64497fcf082defd4a6a0935febce8ed90c7e078
> src/main/java/org/apache/aurora/scheduler/async/TaskScheduler.java 882cdfd9b79b262befb81437cbd9a31a6bc1e40f
>
> Diff: https://reviews.apache.org/r/25872/diff/
>
>
> Testing
> -------
>
> ./gradlew build -Pq
>
>
> Thanks,
>
> Bill Farner
>
>
Re: Review Request 25872: Instrument task scheduling code to give
visibility into failed searches and backoffs.
Posted by Maxim Khutornenko <ma...@apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25872/#review54138
-----------------------------------------------------------
Ship it!
Ship It!
- Maxim Khutornenko
On Sept. 20, 2014, 3:06 a.m., Bill Farner wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25872/
> -----------------------------------------------------------
>
> (Updated Sept. 20, 2014, 3:06 a.m.)
>
>
> Review request for Aurora, Kevin Sweeney and Maxim Khutornenko.
>
>
> Repository: aurora
>
>
> Description
> -------
>
> I'd like to use this data to investigate potential changes to the backoff strategy used in TaskGroups. I have a hunch that tasks are usually either scheduled on the first try, or after the 10th. If that is true, we can make difficult-to-schedule tasks much less expensive.
>
>
> Diffs
> -----
>
> src/main/java/org/apache/aurora/scheduler/async/TaskGroups.java f64497fcf082defd4a6a0935febce8ed90c7e078
> src/main/java/org/apache/aurora/scheduler/async/TaskScheduler.java 882cdfd9b79b262befb81437cbd9a31a6bc1e40f
>
> Diff: https://reviews.apache.org/r/25872/diff/
>
>
> Testing
> -------
>
> ./gradlew build -Pq
>
>
> Thanks,
>
> Bill Farner
>
>