-
Notifications
You must be signed in to change notification settings - Fork 92
/
TODO
30 lines (29 loc) · 1.51 KB
/
TODO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
TODO:
DONE:
* Make it possible to clean up jobs through something like 'crowd cleanup'.
* Think about adding a CLEANUP state, with WorkUnits being dispatched to
workers to clean up a Job's leftover files on S3, instead of the current
in-request cleanup. (see Job#cleanup_assets)
* Right now actions just process. Think about map / process / reduce.
* Basic unit tests.
* Job creation UI. Pick an action, add inputs, set options, go. (not needed)
* Look into Grand Central Dispatch for ideas about how to implement
Job ordering/piping around.
* Finish fleshing out the crowd command.
* Security. HTTP Basic Auth via config.yml,
* Security. S3 authenticated URLs.
* Use the email field to send mail when the Job's gone bad. (bad idea)
* Benchmarking mode -- benchmark every single aspect of a job and report it.
* Pare down the code that the workers are loading to the absolute minimum.
* Documentation. (document inside of config.yml, too)
* Think about having the completion of a work unit respond with another one,
instead of having to make a second request.
* Think about having a failed work unit shuffle off to the end of the queue,
instead of going again right away (and probably being handed back to the
very same daemon).
* Replace work_units_remaining with a percent_complete
* Monitor UI -- see queue of jobs / work units.
* Basic integration tests.
* Unit tests for lib.
* Bump to version 0.1 and release on RubyForge.
* Re-architected to use real Nodes (version 0.2).