| 2013-01-21 02:36:31 utc | randomcamel | is there any reason Dashboard.kill(some_wfid) wouldn't work? it's an FsStorage and it's finding the wfids, but neither #kill, #kill_process, nor #cancel seem to work. |
| 2013-01-21 02:39:47 utc | jmettraux | randomcamel: hello |
| 2013-01-21 02:40:21 utc | randomcamel | heya |
| 2013-01-21 02:40:24 utc | jmettraux | do you have more details? |
| 2013-01-21 02:40:33 utc | jmettraux | you could maybe package that as an issue report |
| 2013-01-21 02:41:13 utc | randomcamel | okay. |
| 2013-01-21 02:42:25 utc | randomcamel | I figured those methods must work in general, so I'm doing something silly, but I'll run a clean case and submit it. |
| 2013-01-21 02:42:58 utc | jmettraux | yes, they are tested |
| 2013-01-21 03:03:23 utc | randomcamel | still working on grokking The Ruote Way, but I feel like it's important that I do. |
| 2013-01-21 03:04:27 utc | jmettraux | same thing for any tool |
| 2013-01-21 03:05:13 utc | jmettraux | just pushed an issue template thing, might help in the future: https://github.com/jmettraux/ruote/blob/master/snips/issue_template.rb |
| 2013-01-21 10:57:40 utc | hartog | jmettraux: it has been accomplished. You can now take https://gist.github.com/4546918 and place it on http://ruote.rubyforge.org/testing_participants.html |
| 2013-01-21 10:57:53 utc | hartog | jmettraux: blessings be with thee ;) |
| 2013-01-21 11:02:30 utc | hartog | wait - closing the channel was *not* the idea |
| 2013-01-21 11:18:26 utc | jmettraux | hartog: many thanks! I'll do that asap |
| 2013-01-21 11:23:54 utc | hartog | jmettraux: np - done with love |
| 2013-01-21 11:24:39 utc | jmettraux | +1 http://otakugangsta.com/post/40783993366 |
| 2013-01-21 19:53:03 utc | randomcamel | trying to separate the dashboard and worker calls does not appear to be a first-class use case. |
| 2013-01-21 19:53:50 utc | randomcamel | ACTION breaks down and reads every page of the docs very carefully =) |
| 2013-01-21 21:26:07 utc | jmettraux | randomcamel: I don't understand what you mean with your statement |
| 2013-01-21 21:26:32 utc | jmettraux | code that uses ruote is generally doing it via the dashboard |
| 2013-01-21 21:26:56 utc | jmettraux | the worker is on its own, it fetches work and performs |
| 2013-01-21 21:27:01 utc | jmettraux | there are no calls for it |
| 2013-01-21 21:37:48 utc | randomcamel | so for a node that is notionally only running worker+participants, I'd still run the worker via the dashboard? |
| 2013-01-21 21:37:59 utc | jmettraux | no |
| 2013-01-21 21:38:08 utc | jmettraux | worker(storage) is sufficient |
| 2013-01-21 21:38:30 utc | jmettraux | there are 3 possibilities |
| 2013-01-21 21:39:00 utc | jmettraux | dashboard(worker(storage)) / worker(storage) / dashboard(storage) |
| 2013-01-21 21:39:19 utc | jmettraux | worker(storage) is for pure work "nodes" |
| 2013-01-21 21:39:34 utc | jmettraux | dashboard(storage) is for steering/querying nodes |
| 2013-01-21 21:42:21 utc | randomcamel | right, but Ruote::Worker.new( Ruote::FsStorage.new('work') ) doesn't appear to wait for work--it exits. |
| 2013-01-21 21:43:12 utc | jmettraux | please #join it |
| 2013-01-21 21:43:30 utc | jmettraux | worker.join |
| 2013-01-21 21:44:45 utc | randomcamel | still exits. I don't trust my PoC, though, I'm going to code something using your issue template just to make sure. |
| 2013-01-21 21:46:02 utc | jmettraux | testing on my side |
| 2013-01-21 21:46:21 utc | jmettraux | exits as well |
| 2013-01-21 21:46:31 utc | jmettraux | let me look into it |
| 2013-01-21 21:46:48 utc | randomcamel | thanks |
| 2013-01-21 21:49:05 utc | jmettraux | you're right, the dashboard calls run on the worker, when the worker is not wrapped, nobody calls run |
| 2013-01-21 21:49:47 utc | jmettraux | https://gist.github.com/4589692 is ok |
| 2013-01-21 21:51:42 utc | jmettraux | https://gist.github.com/4589707 is equivalent, the dashboard can remain unused, it's shallow |
| 2013-01-21 21:59:40 utc | randomcamel | thanks for confirming. that's what I mean by "not a first-class use case". =) it's not like it's a moral flaw, but various things about running ruote in a distributed/multi-node context aren't apparent from the docs. I'll write stuff up once I grok what's going on, though. |
| 2013-01-21 22:01:04 utc | randomcamel | I am excited about the amount of stuff we wouldn't have to implement if we can wire ruote into the system. |
| 2013-01-21 22:01:14 utc | jmettraux | ok :-) |
| 2013-01-21 22:06:49 utc | jmettraux | pointers about how to fix the doc is much appreciated, lessen the friction |
| 2013-01-21 22:07:03 utc | jmettraux | ... are much appreciated... |
| 2013-01-21 23:00:50 utc | jmettraux | GoGoGarrett: hello and welcome to #ruote |