Cannot execute program. fork/exec /tmp/[...]/InteractParser: exec format error #634
Assignees
Labels
Comments
|
Nevermind, I've noticed the Java version is not supported (1.8+). Apologies, seems like they're running multiple versions. |
|
Reopened: the issue persists with Java 1.8. |
|
@GKok I don't think your error has anything to do with your Java version, InteractParser is written in C++. I don't know what the |
|
@GKok is having some issues running some of the TPP tools via X11 forwarding from a HPC, I don't know the source of the error. |
|
@guoci do you think it could be a problem with the CentOS binary deploy? |
|
I've asked the HPC administrators their opinion, and they told me that |
|
That could be the reason. Unfortunately, I don't know if you can change the temporary folder location using FragPipe, you can do that in Philosopher when you run it manually though. |
|
Thanks! I've found this issue, and will try their solution: #306 |
|
That is changing the temp directory of FragPipe. You need to change the temp directory of Philosopher. But, unfortunately, you cannot do that in FragPipe. Best, Fengchao |
|
That's unfortunate. Would it help to re-compile philosopher on the HPC? I guess this would needed to be done prior to every run, as the tmpdir changes every session. |
|
Philosopher is not the issue, the issue is that the FragPipe GUI does not provide users with the option to change the temporary files location. The program has an option that allows you to change the $temp, but the GUI does not expose that for you. |
|
@GKok you can try the following updated version, which will propagate |
|
@guoci Many thanks for your prompt update! This indeed passes the correct tmp directory to Philosopher, and fixes the errors above. However, it now crashes at the same step with a different error. |
|
@GKok does the file |
|
@guoci Yes, it exists and works. |
|
@GKok I don't know what is the cause then. Perphaps you can set the tmpdir to a normal location (outside of |
|
@guoci Thank you for the suggestion. I've tried that, in combination with updating to Philosopher 4.2.1, but unfortunately it throws the exact same error. Googling tells me the error is caused an attempt to modify an executable (InteractParser) while it's running. I'll check with our HPC administrators if they can think of something, and I'll keep trying :-) Once again, thank you all for your assistance. |
|
@GKok I think I might have fixed the issue, please try the following: |
|
@guoci Thanks for the update! This one crashes immediately after initializing the run. I did, however, manage to workaround the error by using the FragPipe 17.2-build 33 you sent me, with Philosopher 4.1.1 (instead of 4.2.1) and deselecting 'generate peptide-level summary' and 'generate protein-level summary'. This way, I get through to the end and get the reports I need. Then, I noticed that the philosopher steps are quite time-consuming. While I'm running on a HPC with many cores available, FragPipe runs every philosopher step sequentially instead of in parallel, ignoring the 'parallelism' setting in the 'Workflow' tab. Do you know if there's another setting that I'm missing, to use the full CPU power available? It would save an enormous amount of time if FragPipe could run the philosoper steps (database/filter/etc.) in e.g., 128 parallel threads. Thanks again :-) |
|
@GKok Please, try a new update again: |
|
@guoci Almost there, but it crashed at the end. |
|
Please try an updated |
|
@guoci This version, in combination with your latest FP build works! And indeed approx. 25% faster than the previous version. Many thanks! :) |
|
Well... That's when using the first 50 files from 2 experiments. I now tried with 250 files from 10 experiments. That crashes again, now at Interstingly, the file it's referring to seems to exist as a tmp-file: And a separate And, another one in a third run: Update: reading the errors more carefully got me thinking, could this be because I was running FragPipe simultaneously in multiple instances on different HPC nodes? Looking at the files above, it seems to be writing some of the temporary files in the mzml directories. If different runs create and delete the same temporary files, this naturally would cause problems. Using the $TMPDIR, all three instances would have different tmp-directories (/scratch/unique-number), but that doesn't help if it's writing in the mzml source folder. |
|
@GKok can you post the full log? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi!
I've used FragPipe on a local Windows machine for optimization of our workflow, but want to run it on a high performance cluster with X11 forwarding for the actual analyses.
Unfortunately, it crashes after a while, with the following set of errors repeated many times:
Machine:
With latests versions of FragPipe, MSFragger and Philosopher.
Any ideas on what's going wrong?
The text was updated successfully, but these errors were encountered: