funannotate
funannotate copied to clipboard
nextgenusfs
OSError: [Errno 28] No space left on device
Hi, An error (OSError: [Errno 28] No space left on device) occurred when running exonerate. I checked my device, it looks like there is plenty of space left.
Filesystem Size Used Avail Use% Mounted on udev 363G 0 363G 0% /dev tmpfs 73G 187M 73G 1% /run /dev/sdg1 902G 656G 201G 77% / tmpfs 363G 0 363G 0% /dev/shm tmpfs 5.0M 0 5.0M 0% /run/lock tmpfs 363G 0 363G 0% /sys/fs/cgroup
Below is my nohup output:
[Apr 27 04:47 PM]: OS: Ubuntu 16.04, 48 cores, ~ 760 GB RAM. Python: 3.7.10
[Apr 27 04:47 PM]: Running funannotate v1.8.7
[Apr 27 04:47 PM]: Skipping CodingQuarry as --organism=other. Pass a weight larger than 0 to run CQ, ie --weights codingquarry:1
[Apr 27 04:47 PM]: Parsed training data, run ab-initio gene predictors as follows:
Program Training-Method
augustus busco
genemark selftraining
glimmerhmm busco
snap busco
[Apr 27 05:07 PM]: Loading genome assembly and parsing soft-masked repetitive sequences
[Apr 27 05:10 PM]: Genome loaded: 28,065 scaffolds; 1,796,331,457 bp; 57.68% repeats masked
[Apr 27 05:10 PM]: Aligning transcript evidence to genome with minimap2
[Apr 27 05:40 PM]: Found 647,976 alignments, wrote GFF3 and Augustus hints to file
[Apr 27 05:40 PM]: Extracting hints from RNA-seq BAM file using bam2hints
[Apr 27 07:40 PM]: Mapping 703,369 proteins to genome using diamond and exonerate
[Apr 27 10:51 PM]: Found 29,021,086 preliminary alignments --> aligning with exonerate
Traceback (most recent call last):
File "/expt/expt_1/clairew/biosoft/conda/miniconda3/envs/funannotate/lib/python3.7/site-packages/funannotate/aux_scripts/funannotate-p2g.py", line 323, in
this seems like it could be transient filespace usage or the files are cleaned up after funannotate aborts. I'd check the input folder size and/or monitor as you run funannotate to see when that occurs. running a df and some du -h during the run in another shell to see if you can track the file size usage.
The exonerate runs are writing to a temp dir p2g_20280073-2d04-4d7a-a2c8-d0a86ce00687 in your example which is cleaned after funannotate ends so you can't really check the size. It looks like you have 29M protein alignments to refine which may be a pretty big result file.
If this is a shared compute is it possible someone else also was writing to same shared disk?
Its unclear which filesystem you are running on relative to your df report above.
Sometimes I run funannotate in a temp dir on scratch partition and then migrate the data back over but your local disk should have more space than your /tmp anyways.
You could potentially reduce the size of your informant proteins and see if that improved your success - are you giving swissprot as the input db or something bigger?
Thanks for your reply. I think you are right. Someone is running a heavy job at the same time on the disk. Also, I have multiple types of input db, including Bam file, closely related protein evidence, swissprot, and transcript evidence, which may produce huge amount of tmp files.
Here is my original script: funannotate predict -i masked_genome.fna --transcript_evidence Trinity-GG.fasta --protein_evidence close_species.faa $FUNANNOTATE_DB/uniprot_sprot.fasta --rna_bam merged.bam --genemark_mode ET --busco_seed_species fly--optimize_augustus --busco_db insecta --organism other --max_intronlen 200000 --name Pam --repeats2evm --cpus 10 --out predict_output
I will rerun funannotate with less protein evidence and see if that works.
Many thanks for your help.