Unicycler failing in covid-19-assembly workflow

Attempting to run usegalaxy.org/u/aun1/h/covid-19-assembly and both Unicycler and Spades are failing with

code 17: slurm_submit_batch_job error: Invalid account or account/partition combination specified
1 Like

Hi @cutsort

This was a cluster issue we have been working on (across all Unicyler/SPAdes runs). It was resolved today. Please try a rerun.

Thanks for reporting the problem!


I’ve been trying to run Unicycler as part of the Sars-CoV-2 genome assembly tutorial and the job has been in que since 11th Aug 2020. I’ve tried a new run aswell and have also run a subsample of the dataset. But, its the 14th of Aug and the jobs are still all on que. Are there any server issues? Please advise what I should do next.


1 Like

Hi @nirvtul

The cluster that runs Unicycler from usegalaxy.org is very busy, with a longer queue than usual. There are not any known technical issues – just more queued jobs than expected.

Please leave your queued jobs queued and those will process as resources become available. If you delete and rerun, the new jobs will end up at the back of the queue again, further extending the wait time.

Our administrator is double-checking why that particular cluster is so busy. If there is some problem uncovered that requires action on your part (rare!), we’ll write back – but most likely the queue will simply start to move again and your existing job(s) will process.

Thanks for reporting the issue and we appreciate your patience. This is a very popular workflow under heavy usage.


Hi jennaj,

Thank you for the clarification. Seems like I’ve already made a mistake…I deleted those old jobs and set new ones with subsampled datasets, thinking that the big datasets would take more time. I’ll stick to your advice and not delete the runs.

Hope the clusters get unclogged…

Best regards

1 Like

@nirvtul No problem! The cluster queue is moving faster again now. Jobs should run a bit quicker but the same advice still applies – leave queued jobs queued for the fastest execution.

In some cases, jobs are scheduled at different queues based on the input size, but not always. This tool is an example exception. At usegalaxy.org, Unicycler always runs on a specific high-memory cluster, and it tends to be in high demand/usage.


Hi @jennaj,

The Unicycler cue is jammed again. Its been about two weeks now. As per your advice, I haven’t deleted the jobs but they’re still on grey status. Hope something can be done soon.


Thanks for reporting the issue @nirvtul – we’ll take a look and get back to you.

Meanwhile, we’d like to review your exact situation, intact.

  1. Startup a rerun of one of your Unicycler jobs – in the same history where the queued jobs are.
  2. Leave the others queued.
  3. After both are done, generate a share link to the history, being sure to check the box to also share datasets in history.

From there, two ways to make sure that we are examining the same account/history. I want to make sure we are looking at the exact same data/jobs. Your choice:

  1. I’ll send you a direct message here. You can reply with:
    • The history share link, and
    • Your usegalaxy.org registered account email address (if different from the email used to register your account here at Galaxy Help).
  2. Alternatively, please send an email from your registered account email address to the galaxy-bugs@lists.galaxyproject.org mailing list. Include the history share link and a link to this post so we can associate the two.

Do not post your account password – even in a private message. And do not include it in your email. Never share that with anyone no matter who they state they are. Administrators will never ask for or need it.


Hi @jennaj,

Many thanks for the speedy response. As per your instructions, I’ve set a new run of the Unicycler job. They are processes 67 and 68. The old ones are processes 57 and 58 and 53 and 54.

Here is the link for the History I’m working on.

Hope the problem gets diagnosed soon.


1 Like

Hi @jennaj,

I’ve set a new run of the Unicycler job and shared a link to the History I’m currently working on in the email. I hope it will be clear to diagnose the problem. Please let me know if I have to supply any other details or if I’m doing anything wrong in the pipeline.

Many thanks.

1 Like