Problem creating Job - Cant select Benchmarks

Hi!

I have problems creating a benchmark job. I uploaded a solver to my space. A subspace contains benchmarks, some of which I uploaded myself, others stem from a uploaded space-xml with references to benchmarks from a public space.
For some reason I cant create a job.

With the the web-interface I have no benchmarks to select. If create a job with everything in the space a message appears with:
http 400 - bad request
Error: no job pairs created for the job. Could not proceed with job submission.

With the command line interface I get:
ERROR: Error communicating with server
I tried several qids, from which I am unsure if they are valid.

Yesterday I got it working using the web interface and I cant see any difference in the steps I'm using now.

Regards,
Florian

The createjob command on the command line tool is still broken (as far as I am aware).

Does the space directly contain any benchmarks? Or are all of the benchmarks in subspaces? Did you used to have a single flat space containing benchmarks? If so, I suspect that you want to create the job to be "all in hierarchy" instead of "all in <space name>". You may be failing into one of the the cases listed in the [slightly out of date] tree given here.  Otherwise, can you give a more detailed description of how you are creating the job with the web interface?

With the steps described in the post you are linking to I was able to start a job.
Thank you!

I have one more question:
In the "solver summary" benchmarks are either summed up in "complete" or "incomplete", but not in "solved", "wrong" or "failed". How can I change this?

> In the "solver summary" benchmarks are either summed up in "complete" or "incomplete",
> but not in "solved", "wrong" or "failed". How can I change this?

I don't even know what these names mean, cf. (unanswered so far)
https://starexec.forumotion.com/t36-what-is-the-semantics-of-subspace-summaries-and-graphs

I guess (I hope) it depends on postprocessors' output, which can be defined by the community, cf.
https://github.com/jwaldmann/ceta-postproc/blob/master/Main.hs

asdfgh wrote:In the "solver summary" benchmarks are either summed up in "complete" or "incomplete", but not in "solved", "wrong" or "failed". How can I change this?

I do not know. For what it is worth, I personally post process my jobs locally using only the "status" as a hint for how to process the job. Statuses are things like "complete", "incomplete", "timeout (cpu)", etc. I do this in order to also get statistics, which are not yet supported.