Local Oasis stuck "Waiting for results"

[Andre_Schleife]

Can the processing get stuck in this state and, if so, how can I cancel this task and either not upload or re-upload?

Thanks for the help!

[mscheidgen]

Unfortunately, if something goes really wrong, everything can happen. There are a few CLI functions you can use. You need to execute them from one of the Oasis containers, e.g. with something like this:

docker exec -ti nomad_oasis_app bash

Then you have a few nomad commands to manipulate uploads. You can “reset” uploads, i.e. put back their processing state:

nomad admin uploads reset --with-entries -- <upload_id>

After this, you should be able to delete them in the UI again.

You can also ls, delete, or process them again (you might need to reset them first though):

nomad admin uploads ls --processing
nomad admin uploads rm -- <upload_id> <upload_id>
nomad admin uploads process -- <upload_id>

See also this bit on the documentation: Operating an OASIS - Documentation.

Let us know, if there are more problems with that.

[Andre_Schleife]

Perfect - this works really well!! (I now prefer it over the web interface )

One followup: I do have now this message when publishing (on our internal Oasis):

Publishing Results in Error "Cannot pack an upload that is packed, or packing"

Dear NOMAD Team, I’m having a problem similar to this one: When I try to publish my upload, the following error appears: Process publish_upload failed: AssertionError: Cannot pack an upload that is packed, or packing. I tried reprocessing the data and managed all the upcoming errors iteratively, but now I’m stuck… Could not find anything helpful on the web, that would give me insight on how to fix this. Could you perhaps have a look at my upload (vkpk-DitQje2Z_zIefdepg) and let me know if I’m…

What/where would I look? I realize the logs may have the info, but they are so massive, that I am not sure where to look or what to look for…

Andre

[mscheidgen]

The unpublished uploads are stored in .volumes/fs/staging/<prefix>/<upload_id>, the published uploads are stored in .volumes/fs/public/<prefix>/<upload_id>. When publishing a temporary file .frozen is created in the staging folder and the system starts to populate the respective public folder. The error message is about the .frozen file. I guess a previous attempt to publish, failed somehow.

You should remove the unfinished public folder and the .frozen file. Then you can reset the processing:

nomad admin uploads reset --success -- <upload_id>

and try to publish again.

If this fails again, please send the latest logs from the nomad_oasis_worker container. The last couple thousand lines should be fine:

docker logs -n 2000 nomad_oasis_worker

Let’s see if we can fix this.

[Andre_Schleife]

Perfect - this solved the problem!!

My last question (for now ):

I am running processing of an upload in a terminal, using the command you gave above:

nomad admin uploads process

But it is “stuck” in:

1 uploads selected, processing ...

For a while now (3 days). Any suggestions for that? Would nomad_oasis_worker log output help with that, too?

[mscheidgen]

Yes, the logs will help. I assume it is running into a problem, where it is unable to report an error or problem. With the logs, let us also know, which version (e.g. image tag) you are using.