core dump in goldrush-edit about goldrush HOT 7 CLOSED

Hi Lauren,

Terribly sorry for the late reply.
ulimit -u 256000 didn't change anything, but downgrading btlib to v1.4.10 fixed it!
The test run worked out fine, and I'm running it on my real data right now.
Big thanks for the advice!

from goldrush.

Hi @abs-yy,

Thanks for the report! Could you let us know what type of machine you are running the tests on (including OS version, RAM, number of threads, etc)?

Thank you for your interest in GoldRush!
Lauren

from goldrush.

Hi Lauren,

I'm using our institute cluster server, qsh into a high memory node and running goldrush.
I had access to 36 threads and and probably ~1Tb RAM.

here's some info on the node I've used. If there's anything else I can provide please tell me!

$ cat /etc/centos-release
Red Hat Enterprise Linux Server release 7.2 (Maipo)

$ uname -a
Linux t01 3.10.0-327.36.3.el7.x86_64 #1 SMP Mon Oct 24 16:09:20 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux

from goldrush.

When I run goldrush with a single core, I get this python3 error as well.

2023-03-08 12:40:55][INFO] Polishing batches...
*** Error in `python3': corrupted size vs. prev_size: 0x00002b98d1ea83e0 ***
======= Backtrace: =========
/lib64/libc.so.6(+0x80f17)[0x2b98cb4fdf17]
/lib64/libc.so.6(+0x82415)[0x2b98cb4ff415]
/lib64/libc.so.6(+0x839ac)[0x2b98cb5009ac]
/lib64/libc.so.6(realloc+0x1d2)[0x2b98cb5029b2]
/lib64/libc.so.6(getdelim+0x10b)[0x2b98cb4ec8ab]
${HOME}/anaconda3/envs/test/lib/python3.9/site-packages/btllib/_btllib.so(_ZN6btllib9SeqReader13readline_fileERNS_7CStringEP8_IO_FILE+0x16)[0x2b98d3163146]
${HOME}/anaconda3/envs/test/lib/python3.9/site-packages/btllib/_btllib.so(_ZN6btllib9SeqReader20readline_file_appendERNS_7CStringEP8_IO_FILE+0x1c)[0x2b98d316316c]
${HOME}/anaconda3/envs/test/lib/python3.9/site-packages/btllib/_btllib.so(+0xf7c26)[0x2b98d3163c26]
${HOME}/anaconda3/envs/test/lib/python3.9/site-packages/btllib/_btllib.so(+0xfb538)[0x2b98d3167538]
${HOME}/anaconda3/envs/test/lib/python3.9/site-packages/btllib/../../../libstdc++.so.6(+0xcda93)[0x2b98d32e9a93]
/lib64/libpthread.so.0(+0x7dd5)[0x2b98ca957dd5]
/lib64/libc.so.6(clone+0x6d)[0x2b98cb57b02d]
======= Memory map: ========
## A whole lot of memory mapping data below

from goldrush.

Hi @abs-yy,

Thanks for your system info - I was wondering about the memory, but you clearly have plenty of threads and RAM to work with. We also run most of our tests jobs on centOS machines here, so the OS shouldn't be an issue.

Could you try running ulimit -u 256000 prior to your GoldRush command? While your error message isn't exactly the same, other users have reported issues at this stage of the polishing module, and that command helped. So, I think it's worth a try.

from goldrush.

Hi @abs-yy,

Another thought - Could you check what version of btllib is installed? If you have v1.5.0, it has come to our attention that there is an issue with this latest version of btllib. In that case, downgrading to v1.4.10 should help.

from goldrush.

Excellent, I'm so glad that fixed it!
As a side note, we are working on a fix for that btllib issue (seen in v1.5.0) - we should have a new release later this week.

from goldrush.