search

crux-toolkit / crux-toolkit

http://crux.ms/
Other
32 stars 26 forks source link

seg fault from tide-index #338

Closed CharlesEGrant closed 8 years ago

CharlesEGrant commented 8 years ago

I am getting a seg fault from this command line:

/net/noble/vol1/home/noble/proj/crux/trunk/release/src/crux tide-index --mods-spec 1KR+14.0157,1MLPDHY+15.9949,1LR+27.9949,1KR+28.0314,1K+42.0470,1KST+42.0106,1L+68.0261,1STY+79.9663,1L+114.0429 --nterm-peptide-mods-spec 1MSAV+42.0106,1M+58.0055 --max-mods 3 --enzyme no-enzyme --overwrite T --peptide-list T --decoy-format none --min-length 7 --min-mass 700 --output-dir maxMods=3.perMod=1 ../../../data/uniprot/histones.fa histones

The input file and log file are attached.

Reported by: wsnoble

CharlesEGrant commented 8 years ago

I think it might have been caused by running out of disk space. I tried running it on a machine with ~14G free which wasn't enough. Looking at pyrrolysine, Filesystem Size Used Avail Use% Mounted on /dev/mapper/sysvg-rootvol 20G 19G 18M 100% / I summed up the temporary modified_peptides_partial files in /tmp and they only add up to 7.2G.

Original comment by: kaipot

CharlesEGrant commented 8 years ago

Any chance this problem could be detected and reported as an error, along with some kind of error message, rather than leading to a seg fault?

On Wed, Jul 6, 2016 at 12:54 PM, Kaipo kaipot@users.sf.net wrote:

I think it might have been caused by running out of disk space. I tried running it on a machine with ~14G free which wasn't enough. Looking at pyrrolysine, Filesystem Size Used Avail Use% Mounted on /dev/mapper/sysvg-rootvol 20G 19G 18M 100% / I summed up the temporary modified_peptides_partial files in /tmp and they

only add up to 7.2G.

Status: open Milestone: Crux v2.0 Labels: High priority Created: Wed Jul 06, 2016 04:48 PM UTC by William S Noble Last Updated: Wed Jul 06, 2016 04:48 PM UTC Owner: Kaipo Attachments:

I am getting a seg fault from this command line:

/net/noble/vol1/home/noble/proj/crux/trunk/release/src/crux tide-index --mods-spec 1KR+14.0157,1MLPDHY+15.9949,1LR+27.9949,1KR+28.0314,1K+42.0470,1KST+42.0106,1L+68.0261,1STY+79.9663,1L+114.0429 --nterm-peptide-mods-spec 1MSAV+42.0106,1M+58.0055 --max-mods 3 --enzyme no-enzyme --overwrite T --peptide-list T --decoy-format none --min-length 7 --min-mass 700 --output-dir maxMods=3.perMod=1 ../../../data/uniprot/histones.fa histones

The input file and log file are attached.

Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/cruxtoolkit/issues/411/

To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/

Original comment by: wsnoble

CharlesEGrant commented 8 years ago

It failed on my machine with a (not very user-friendly) error about a writer failing, rather than a segmentation fault, not sure why the difference. Anyway, it should be straightforward to have a generic i/o error message, though I'm not too sure about distinguishing between the different types of i/o errors.

Original comment by: kaipot

CharlesEGrant commented 8 years ago

I don't think it's critical that we distinguish between different types of errors. As long as the error gets caught and indicates that it has something to do with I/O, we should be fine. On pyrrolysine, compiled from the current trunk, I get

INFO: Reading proteins INFO: Computing modified peptides... INFO: Opening 2048 files for modifications. Segmentation fault (core dumped)

Original comment by: wsnoble

CharlesEGrant commented 8 years ago

Here is a patch that shows an error on failed writing: INFO: Reading histones.fa and computing unmodified peptides... INFO: Reading proteins INFO: Computing modified peptides... INFO: Opening 2048 files for modifications. FATAL: I/O error writing modifications

Could you give it a try and make sure it works for you?

Original comment by: kaipot

CharlesEGrant commented 8 years ago

I have verified that this patch works as expected on Linux and Mac OS. You can go ahead and check it in and close this ticket.

Original comment by: wsnoble

CharlesEGrant commented 8 years ago

Original comment by: kaipot

CharlesEGrant commented 8 years ago

Fix committed in r17069

Original comment by: kaipot