More than 1 week job waits for RNA STAR

Hello. Twice in the month of July I’ve been unable to successfully run RNA STAR 2.7.8a on the usegalaxy.org server. During the first week of July I had this issue and deleted my job after a week of waiting. I’m currently attempting another job, which has remained queued for the last 6 days. I’m curious if this is expected behavior for this application. Thank you and I apologize for the bother.

1 Like

Dear @pad59,
Unfortunately, it is. Depending on the organism, your input data, some settings, and the day, STAR can take very long.

For example:

  • Mapping to human in comparison to cyanobacteria takes longer.
  • The deeper you sequence the longer it will take to map your data.
  • Use 2-pass mapping for more sensitive novel splice junction discovery activating this function also takes longer.
  • If you use also a non-standard genome/transcriptome, which is not supported by Galaxy, then STAR takes especially longer, because it has to first generate an index file and then runs the mapping.
  • It could be that you have used Galaxy on days, when a workshop, conference or training was given. During those peak times, jobs can take a bit longer.

I hope I could clarify, why your job might have taken so long. Maybe to speed up your process here are few things you can do:

  • Re-run your job, like you have done already.
  • Check a few settings for STAR to speed up the process (like the 2-pass mode, if you have selected it).
  • If you have used a non-standard genome/transcriptome, then maybe it is better to first run RNA STAR locally once to generate the index file, which can then be provided for Galaxy.

I hope I could help.

Have a good week and best wishes,
Florian

1 Like

Thanks @Flow for the detailed help on running the tool. I will add that we had a cluster issue on usegalaxy.org that was allowing only 1 RNA STAR job to execute at a time, which was likely the reason the second job was not getting to the running stage. That issue has been fixed and we’re working through a backlog of jobs, so @pad59 if your job does not run within the next 24 hours, please let us know.

1 Like

Hi Nate,

I have also been having problems with some jobs for >1 week (Trinity but only on Org). They fail once they move out of the queued state. Jen got back to me when I reported the error, which is likely a configuration problem on the server or the remote cluster and she said that RNA STAR may also possibly be affected by this problem. I see you mention the cluster issue has been solved, but I had a Trinity fail last night again (same error message as I have been getting all week). Do you know if this issue has been figured out?

1 Like

Thank you Nate and Florian! The cluster issue Nate mentioned may be relevant to my problem since I am not doing any of the computationally-intensive tasks that Florian listed, and have never had more than a couple-day wait for my RNA STAR jobs over the last several years.

To update @nate, it’s now two days since your cluster issue fix and my job is still queued (since July 18 overall). I’ll keep it queued for now unless you suggest otherwise. Thank you again for your help.

1 Like

Hi @R_R_R and everyone else reading:

Jobs at UseGalaxy.org that failed with a message like the one below were due to a server configuration problem that is now resolved. The correction went into effect later in the day on August 6, 2021. Any jobs that failed with this error should be rerun.

Invalid qos specification

Thanks!

Hi @pad59 and others reading who may have also run into the stalled job problem during this same time frame

Three weeks is much too long of a wait. Unless those have started up by now, you’ll need to rerun them.

Thanks for your patience, and the updates! Very helpful

Thanks @jennaj , I was still getting the same error earlier today, but I am starting everything from scratch and running again.

1 Like

Thank you @jennaj. I’m writing to update you on my issue: I deleted and reran my RNA STAR job as you suggested but once again I’m getting a surprising long wait: it’s been 7 days and my job is still queued. I’ve never gotten any sort of error message like @R_R_R did–I simply have this long wait. Over last few years the types of jobs I run typically queue for a day or so at most.

1 Like

@pad59 I’m going to send you a direct message – would like to review your account and these runs to help figure out what the root issue is. A week is too long, too – some jobs should have started up by now.