Hi @antoinedao,
Thanks!
Unfortunately the error is still there.
I tried a new case/job. It took a couple of hours to finish (about 90K sensors). But when i connect the results to the Load Assets component it returns the 'nasty" error (1. Unable to download file).
I need to update something on my side?
-A.
Thanks again for the feedback. I’m going to look into this further today and get back to you. Sorry about the delays this is causing 
No worries @antoinedao,
I noticed the slightly different error now and before.
The before was “Unable to download results”
and now “Unable to download file”.
Maybe this means something.
-A.
@ayezioro you helped me uncover a bug in our workflow execution layer! Thank you! I have fixed it and tested it against the files you were using and am happy to say that i works!
Please give it one last try and rejoice with me when you eventually manage to download your results from our platform 
Glad to help @antoinedao!
I need to submit the job again or i can i download the existing results?
I’m asking because i’m still having the error.
Thanks,
-A.
Small update,
Now it seems to work. No error reported. But … Null results at the output items of the Load Assets component. Like so:
You will need to run it again unfortunately 
Hi @mostapha,
Just sent it. This is the link.
Thanks,
-A.
p.s. the job is finished. Still, as in the preview post, the outputs show Null values.
It can be caused because my internet connection at home. I’ll check again when i’ll be at the university.
Hi,
Another update.
After a while, the Load Assets component turned red and complained with this error (that we already know):
- Unable to download file
But … checking on my C:\Users\USERNAME\AppData\Local\Temp\Pollination\ folder, i see there this one (from today) jzxy2edt.p43.
Its size is huge 8.1Gb and has 32 *.ill files and sun-up-hours.txt and grids_info.json files.
So basically something good is happening but is not finished (or something else).
I’ll report back from the university when i get there, unless you have some words of wisdom in the meantime.
-A.
I don’t have an insight to offer here but I can confirm that it also fails on my side.
If @antoinedao can confirm that the ZIP file is created then it is an issue on the Grasshopper plugin side.
cc: @mingbo
I can confirm that we’re still encountering issues zipping the files on the cloud end… I have one last trick in my pocket to try and after that we will probably have to consider redesigning our zipping strategy (or encourage ambitious members to not run urban scale projects quite yet )
I’ll push my changes and report back here where they are ready to try out.
That’s not an [ideal] option and I hope we don’t get there! 
But we have two more options to try if this last trick also doesn’t work:
Hi,
My forecast is that the cases using large numbers of sensors will be on the rise. Sometimes because it is possible to do (even though is not necessary) and sometimes because the cloud option opens doors that before were closed.
Either case a proper solution should be provided or the limits need to be set. Right now i think the hurdle is more technical than any other reason. The simulation itself is finished, so for me will be hard to believe that getting the results back is not solvable.
My feeling is that even for you (@mostapha, @antoinedao) is hard to foresee the limits of these possible cases (maybe even millions of sensors), but i’m sure a proper and satisfying solution is on the way!!
Thanks for the efforts. I understand it is not easy.
-A.
Hi Abraham!
You’re right and thank you for being understanding. This is one of those issues that is hard to debug because it’s really hard to recreate it on our side. With the current fixes, if we run the last step in isolation it runs fine but if you run the study from scratch we face the issue that you mentioned. @antoinedao has done a great job pin-pointing the issue and we have at least two other options to try so yes the solution is coming! 
For this specific case another option is to add a post-processing step to the recipe so you don’t have to download the raw annual results. Similar to annual daylight recipe. I know that @chris has been thinking about adding that extra step. What’s the post-processing that you need to do here @ayezioro? What are you trying to visualize? Total radiation over all the timesteps? Or do you need the hourly data for each sensor?
If the post-processing needs a lot of flexibility we can provide the post-processing as its own recipe that reuses the results of a previous run. This needs extra development on the backend so it will take some time.
This is the price that we have to pay for providing “freedom and flexibility for real world complexity”. I think these cases are real world problems and that’s why we started Pollination in the first place. To make these level of complexity accessible to everyone.
When @asisnath was presenting his work during the previous meetup I couldn’t help to think that a solution that was only available to @chris about 5 years ago - for modeling those comfort maps and was very painful to set-up - is now available to everyone who will use Pollination! That’s why we are building this and we know that the process will have ups and downs. Thank you for helping us to get 
to the"Plateau of Productivity" sooner than later! 
Hi @mostapha,
For now i need the total over the period. But i want to filter them according to different min/max values. It is intended to asks questions regarding BIPV potential. So for that i need the numbers altogether with the visuals.
The flexibility in post-processing is important, so i appreciate @chris’s efforts in this direction.
As for my specific case, i’m absolutely confident you’ll get there … as always you do. Just hope my case is enough general that can help find a comprehensive solution.
Keep the good work and efforts. A big community is behind you.
-A.
hi,
Just to report that no success either with the University internet.
This time even those results (ill files) i mentioned are not downloaded.
-A.
Hi all
I also struggle with the download
First i have difficulties to handle this setting in a way, not to have the previous setting files installed again in a sub folder of my choice before I can install the desired one.
The provided sub folders are not really of any use. Having 20 of these encrypted ones in a project file I don’t how to figure out the content?
In this case the files are already downloaded to the folder except for the visual report. Would be helpful to know which one is failing.
I was going to suggest you using several download components to download these resources. That way we will know which one is causing the issue.
Or maybe make both ways accessible, the existing one and single components?
for the Check Job Status component I have the workaround to disable it before setting a new Folder. But with a “sense making” Folder structure in general, it may not necessary to change something about it?
