abyss-pe failure after files read

[aquat...@g.hmc.edu]

Hi there,

I am having trouble with abyss-pe failing during a run.  Perhaps this is a memory issue? (Although I am working on a 64 core processor with 512 MB RAM). Or perhaps a writing permissions issue?

Thanks in advance,

Andrea

Code:

nohup ~/abyss-pe np=10 k=64 name=SAF099abyss in='SAF099.1.fq SAF099.2.fq'

Here is the begin and end of the nohup.out file:

/usr/bin/mpirun -np 10 ABYSS-P -k64 -q3    --coverage-hist=coverage.hist -s abyss-bubbles.fa  -o abyss-1.fa SAF099.1.fq SAF099.2.fq 

ABySS 2.0.2

ABYSS-P -k64 -q3 --coverage-hist=coverage.hist -s abyss-bubbles.fa -o abyss-1.fa SAF099.1.fq SAF099.2.fq

Running on 10 processors

8: Running on host purves

9: Running on host purves

0: Running on host purves

0: Reading `SAF099.1.fq'...

1: Running on host purves

1: Reading `SAF099.2.fq'...

........................................

Pruned 3273 k-mer in 2030 tips.

Pruning tips shorter than 4 bp...

7: Pruned 259 tips.

4: Pruned 267 tips.

2: Pruned 277 tips.

10: Pruned 276 tips.

3: Pruned 280 tips.

11: Pruned 281 tips.

6: Pruned 255 tips.

8: Pruned 254 tips.

5: Pruned 295 tips.

0: Pruned 271 tips.

1: Pruned 290 tips.

9: Pruned 291 tips.

Pruned 9315 k-mer in 3296 tips.

Pruning tips shorter than 8 bp...

/data/mcfadden/aquattrini/PROGRAMS/abyss-2.0.2/bin/bin/abyss-pe:501: recipe for target 'abyss-1.fa' failed

make: *** [abyss-1.fa] Error 1

[Shaun Jackman]

Hi, Andrea. I'm afraid I don't see an error message in there that suggests why your job failed. If you're running on a HPC scheduler, my best guess would be that exceed either your memory or time limits and the job was killed by the scheduler. To trouble shoot, you'll want to find out why the process `ABYSS-P` died, was it killed by a signal, and what was its exit status. It's odd that there's no error message from `mpirun` as I would expect. Can you please post the entire log using gist.github.com ?

Cheers,

Shaun

[aquat...@g.hmc.edu]

I added the file to gist. I was able to run abyss on the same files again using 8 cores. Seems like a memory issue.  I also just ran into another issue and posted the file there as well.  

[Ben Vandervalk]

Hi Andrea,

Can you post the link to your gist?

I don't think changing the number of cores would change the memory usage much.  (Regardless, it is loading the full k-mer set from the reads into memory.)

- Ben

--
You received this message because you are subscribed to the Google Groups "ABySS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to abyss-users+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[aquat...@g.hmc.edu]

Oops...

https://gist.github.com/quattrinia/b7bc0e157ec48fa6f3db3805b9ff1a02

To unsubscribe from this group and stop receiving emails from it, send an email to abyss-users...@googlegroups.com.