Improve documentation for scheduler states #1498
Conversation
| @@ -2,7 +2,7 @@ | |||
| # | |||
|
|
|||
| # You can set these variables from the command line. | |||
| SPHINXOPTS = | |||
| SPHINXOPTS = -j4 | |||
There was a problem hiding this comment.
This will build docs in parallel.
|
|
||
| * ``{'task': (inc, 1)}``: a tuple satisfying the dask graph protocol. This |
There was a problem hiding this comment.
Putting 'task' there seemed to be a bug.
|
|
||
| A key's priority is only used to break ties, when many keys are being | ||
| considered for execution. The priority does *not* determine running order, |
There was a problem hiding this comment.
"The priority does not determine running order" sounds wrong to me, otherwise what would it be used for? The goal is to start computing some tasks earlier than others...
There was a problem hiding this comment.
Yes, I suspect that my intent here was to say that priority does not fully determine running order. Many other concerns can take precedence. This probably isn't necessary to say though.
docs/source/scheduling-state.rst
Outdated
| processing → erred occupancy, idle, used_resources | ||
| processing → released occupancy, idle, used_resources | ||
| memory → released worker_bytes | ||
| memory → forgotten worker_bytes |
There was a problem hiding this comment.
The task in tasks itself is also forgotten if you want to include this
There was a problem hiding this comment.
I am actually a bit surprised that the right column of this table is not more dense. I would have expected more worker and task state to be affected. Ah, I see that you are only listing worker state, never mind.
|
In general this looks great to me. I agree with all of the wording changes. Thank you for going through this. |
pitrou
added
enhancement
pitrou
changed the title
|
I'm +1 on this PR and happy to merge any time. |
|
Cool, thank you :-) |
|
Thank you! This seems much clearer to me than before.
…On Thu, Oct 26, 2017 at 7:26 AM, Antoine Pitrou ***@***.***> wrote:
Merged #1498 <#1498>.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#1498 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AASszG4u08xW4lLpt5wrQQS6D8AZPrN1ks5swGxagaJpZM4QEzPE>
.
|
This categorizes the state variable in three groups, adds a table of where tasks appear depending on their state, another table listing worker state changes in each task transition, and improves some wording choices in my POV :-)