zooniverse / old-weather

Old Weather aims to recover worldwide weather observations to help improve climate model predictions.
http://oldweather.org
Apache License 2.0
0 stars 3 forks source link

Random "Timeout" or "Hang" issues when loading new subject image (aka the Empty Subject Sets issue) #82

Closed shaunanoordin closed 9 years ago

shaunanoordin commented 9 years ago

Problem:

zooniverse-oldweather-20151120-bug-report

Analysis:

Notes:

zooniverse-oldweather-20151120-bug-report.txt

camallen commented 9 years ago

pretty sure Roger debugged this that the default non-logged in queues weren't built due to the way we imported the data...

rogerhutchings commented 9 years ago

I did, and I edited the subject sets (tweaked set names and reverted), but it looks like it's still an issue

camallen commented 9 years ago

did the names actually save?

rogerhutchings commented 9 years ago

yeah

camallen commented 9 years ago

then it should have redone the q's, https://github.com/zooniverse/Panoptes/blob/69a96b3d4144adf9d162caab07169740b35a5e70/app/controllers/api/v1/subject_sets_controller.rb#L15

camallen commented 9 years ago

Is your workflow grouped? If not then it won't....

rogerhutchings commented 9 years ago

Not as far as I know, it's just grabbing subjects from a set

camallen commented 9 years ago

your workflow has to be grouped if you're using subject set groups. same as annotate..without this the grouping configs don't work.

rogerhutchings commented 9 years ago

They're not groups of sets, just sets

camallen commented 9 years ago

ok - each subject set is considered a group for selection. annotate works this way, check the grouped setting on the workflow. if you want grouped (subject_set) selection to work correctly (including building the non-logged in queues) this has to be set.

rogerhutchings commented 9 years ago

Okay - I've set this, and and edited the sets again, but I'm still getting empty sets. Would you mind having a look @camallen?

shaunanoordin commented 9 years ago

Please let me know if I can help look into the sets or etc - I'm currently running some tests to see the % of getting an empty set in each of the ships, just in case that data might help the debug.

shaunanoordin commented 9 years ago

Update: whatever's been happening in the past 10 minutes, the subject sets have gone from empty 80% of the time to the full 20 randomly-picked subjects in a set, all the time.

Tested for Eastwind 1946. (https://panoptes.zooniverse.org/api/subjects?sort=queued&workflow_id=886&page=1&page_size=20&subject_set_id=2529) Nice!

camallen commented 9 years ago

not sure what's up here..probably won't have time to look today..when is it needed?

camallen commented 9 years ago

please close if this isn't an issue anymore

rogerhutchings commented 9 years ago

Okay, I've checked all of the sets - they seem to be okay, apart from Bear 1899 (subject set id 2493). Doesn't need to be today, but can you take a look on Monday?

rogerhutchings commented 9 years ago

I take it back, I think it's okay :D Thanks all