search

alekseyzimin / masurca

GNU General Public License v3.0
243 stars 35 forks source link

Restarting an assembly - can't open overlap.sh #205

Closed sjfleck closed 3 years ago

sjfleck commented 3 years ago

I'm doing a couple assemblies using MaSuRCA-3.4.3. I'm also only using only trimmed illumina paired end reads (150 bp). The first assembly finished in about 34 hours and was ~95 Mbp. The second assembly is estimated by Kmergenie to be ~185 Mbp and ~245 Mbp by MaSuRCA. I'm working on the SLURM cluster at my university where there is a 72 hour time limit for jobs to run. This larger assembly hit the time limit at 72 hours and was canceled automatically at that point. Since it wasn't a typical error, I thought that I could simply restart it using the recommendations for errors due to running out of space:

"if you noticed that CABOG failed due to lack of disk space, then, after freeing some space, simply run $/install_path/MaSuRCA-X.X.X/bin/masurca config.txt and execute assemble.sh"

Of course I didn't need to free up space, but when I restarted the assembly, I got this error very quickly:

No need to run meryl for OBT (OBT is disabled). No need to run meryl for OVL (genome.nmers.ovl.fasta exists). runCA failed. Stack trace: at /full/path/MaSuRCA-3.4.3/bin/../CA8/Linux-amd64/bin/runCA line 1613. main::caFailure("can't open '/panfs/panfs.cbls.ccr.buffalo.edu/scratch/grp-vaa"..., undef) called at /full/path/MaSuRCA-3.4.3/bin/../CA8/Linux-amd64/bin/runCA line 3681 main::createOverlapJobs("normal") called at /full/path/MaSuRCA-3.4.3/bin/../CA8/Linux-amd64/bin/runCA line 6525 Failure message: can't open '/full/path/CA/1-overlapper/overlap.sh'

I looked in the 1-overlapper directory and I did not see a file called overlap.sh. Any insight into what I need to do to restart the assembly would be greatly appreciated. Thank you

sjfleck commented 3 years ago

Hi all, the issue was strictly one of permissions. I typically submit jobs to our SLURM cluster with the command:

sbatch assembly.sh

BUT, this new directory that I'm working in need some more instructions to run MaSuRCA. When I run:

sg "my-groups-name" "sbatch G42_assemble2.sh"

the program begins running just fine. I should have known that this was the issue because I initially had to run MaSuRCA with this sg command. I simply forgot to submit it this way when I needed to resume the assembly. It turns out that MaSuRCA is trying to create new files that are made and owned by me instead of my group, which is against the rules in this directory. I'm glad that this was an easy fix. I hope this helps if anyone else gets a similar issue.