Webinterface "Resubmit from start" on log

[kimble]

Hi,

I'm not sure if this is a bug but the 'Resubmit from start', available as a right click on the webinterface finished output doesn't seem to run the entire workflow on a file, perhaps just the node it failed at again.

I have a workflow with a conditional step before the encode that checks for a file contents. If this process fails, but then I fix the conditional step file so it should pass then a "resubmit from start" still fails on the same issue. As if it's not reflowing through the conditional step again and getting an updated variable.

[emcodem]

Well my first guess would be that your workflow has mutliple starting points and the one that is taken is kind of chosen randomly.
You can more or less currently only control this by making sure your workflow has only one starting processor...
If you share your workflow i might be able to tell you more about it

[admin]

It's not a bug, it's a feature No really, it is. Because:
FFAStrans won't know the difference between changing only some part of the workflow and fundamentally changing it. Imagine if you actually removed the node where the error occured, or remove the starting node for the job. FFAStrans would just have to pick another starting point, or some other node and the whole resubmit might be totally invalid. So FFAStrans will always cache the used workflows at the start of a job. And it's using that when resubmitting. So the use case for resubmitting is not for changed workflows but rather mishaps outside of the control of FFAStrans. At least that is the current design.

-steinar

[emcodem]

oh so forget what i said, steinar is right, but that only counts when you resubmit in ffastrans application. Resubmit in webinterface would probably work for you and consider the changed value of your conditional.

[kimble]

Well my first guess would be that your workflow has mutliple starting points and the one that is taken is kind of chosen randomly.
You can more or less currently only control this by making sure your workflow has only one starting processor...
If you share your workflow i might be able to tell you more about it

Ok, that would explain it. I've separated them out now. Thank you.

[emcodem]

Funny because according to your description it sounded more like the issue that steinar pointed out...
Anyway, good to know how you solved your stuff so we know what to work on and what to document better ;D