strange segfault about minia

[ekg]

I should note that this machine has 256G of RAM and I don't appear to have any limitations as far as disk space.

[rchikhi]

Hi Erik!

Curious, and that's definitely not tied to the number of fastq files, as at this stage minia only considers the counted kmers and no longer the input files. That's the first time I see this stage fail, it could be due to some special unhandled case in the graph structure.

I'd be happy to assist with debugging.

1. Can the data be sent by any chance?
2. Does it complete with a different kmer size?

[ekg]

It does complete with a shorter kmer size (41) and the same abundance. I would love to share the data but I will need to ask my collaborators. I will send you an email if I can share. I definitely understand that's the only way to really resolve this problem. Interestingly the k=41 assembly has very few contigs (in the order of a few hundred kb) while the unitig set is several hundred mb. Also this data is a little strange, it's going to consist of many small fragments by library design (GBS/RadSeq).

[rchikhi]

I see. Well, I'll keep an eye for that email, otherwise, just let me know if you encounter that bug again in another dataset. I'd like to get a sense of whether this is a one-of-a-kind thing.

[Mozart1776]

Hello,
I am also experiencing a segmentation fault of minia at the Iterating DSK partitions step at k41 (see the log file below). I am using a computer cluster with nearly 2TB of RAM and a large disk space. Minia runs without an issue with the test GATB dataset. Any help?
Thanks!

(2018-08-13 22:18:53) GATB-pipeline starting
(2018-08-13 22:18:53) Command line: /home/pavel/bin/gatb-minia-pipeline/gatb -1 /home/pavel/Run_1837_Tyrophagus_putrescentiae_Mex/Sample_79070/2a-removeDuplicates_clumpify/79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_Results/79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.fastq.gz -2 /home/pavel/Run_1837_Tyrophagus_putrescentiae_Mex/Sample_79070/2a-removeDuplicates_clumpify/79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_Results/79070_TGACCA_S1_L003_R2_001_bbduk_no_adapters_deduplicated.fastq.gz -o 79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.assembly

(2018-08-13 22:18:53) Setting maximum kmer length to: 151 bp
(2018-08-13 22:18:53) Multi-k values and cutoffs: [(21, 2), (41, 2), (61, 2), (81, 2), (101, 2), (121, 2), (141, 2)]

(2018-08-13 22:18:53) Minia assembling at k=21 min_abundance=2
(2018-08-13 22:18:53) Execution of 'minia/minia'. Command line:
/home/pavel/bin/gatb-minia-pipeline/tools/memused /home/pavel/bin/gatb-minia-pipeline/minia/minia -in 79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.assembly.list_reads -kmer-size 21 -abundance-min 2 -out 79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.assembly_k21
(2018-08-14 03:36:01) Finished Minia k=21

(2018-08-14 03:36:01) Minia assembling at k=41 min_abundance=2
(2018-08-14 03:36:01) Execution of 'minia/minia'. Command line:
/home/pavel/bin/gatb-minia-pipeline/tools/memused /home/pavel/bin/gatb-minia-pipeline/minia/minia -in 79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.assembly.list_reads -kmer-size 41 -abundance-min 2 -out 79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.assembly_k41
(2018-08-14 05:06:33) Execution of 'minia/minia' failed. Command line:
/home/pavel/bin/gatb-minia-pipeline/tools/memused /home/pavel/bin/gatb-minia-pipeline/minia/minia -in 79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.assembly.list_reads -kmer-size 41 -abundance-min 2 -out 79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.assembly_k41
pavel@sagarana:~/Run_1837_Tyrophagus_putrescentiae_Mex/Sample_79070/5d-gatb-pipeline/79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated> cat 0-gatb_79070_TGACCA_S1_L003_R1_001_bbduk_no_adapters_deduplicated.pbs.e108512
[Approximating frequencies of minimizers ] 100 % elapsed: 0 min 48 sec remaining: 0 min 0 sec cpu: 99.9 % mem: [ 20, 20, 20] MB
[DSK: nb solid kmers found : 595537654 ] 212 % elapsed: 71 min 3 sec remaining: 0 min 0 sec cpu: 1813.0 % mem: [7845, 8032, 8050] MB
[Iterating DSK partitions ] 99.7 % elapsed: 21 min 6 sec remaining: 0 min 4 sec
[Building BooPHF] 100 % elapsed: 0 min 11 sec remaining: 0 min 0 sec
[removing tips, pass 1 ] 100 % elapsed: 8 min 49 sec remaining: 0 min 0 sec cpu: 25282.3 % mem: [28646, 28646, 33965] MB
[removing tips, pass 2 ] 100 % elapsed: 0 min 39 sec remaining: 0 min 0 sec cpu: 24498.1 % mem: [24143, 28675, 33965] MB
[removing tips, pass 3 ] 100 % elapsed: 0 min 21 sec remaining: 0 min 0 sec cpu: 7850.8 % mem: [24082, 24134, 33965] MB
[removing tips, pass 4 ] 100 % elapsed: 0 min 22 sec remaining: 0 min 0 sec cpu: 8396.4 % mem: [24082, 24082, 33965] MB
[removing tips, pass 5 ] 100 % elapsed: 0 min 21 sec remaining: 0 min 0 sec cpu: 7820.0 % mem: [24082, 24082, 33965] MB
[removing bulges, pass 1 ] 100 % elapsed: 10 min 29 sec remaining: 0 min 0 sec cpu: 22194.7 % mem: [24194, 24194, 33965] MB
[removing bulges, pass 2 ] 100 % elapsed: 9 min 7 sec remaining: 0 min 0 sec cpu: 23147.3 % mem: [23158, 24203, 33965] MB
[removing bulges, pass 3 ] 100 % elapsed: 9 min 10 sec remaining: 0 min 0 sec cpu: 23453.0 % mem: [22105, 23163, 33965] MB
[removing bulges, pass 4 ] 100 % elapsed: 9 min 12 sec remaining: 0 min 0 sec cpu: 23583.9 % mem: [21399, 22098, 33965] MB
[removing bulges, pass 5 ] 100 % elapsed: 9 min 44 sec remaining: 0 min 0 sec cpu: 23581.3 % mem: [20617, 21390, 33965] MB
[removing ec, pass 1 ] 100 % elapsed: 5 min 10 sec remaining: 0 min 0 sec cpu: 24824.4 % mem: [21343, 21343, 33965] MB
[removing ec, pass 2 ] 100 % elapsed: 1 min 4 sec remaining: 0 min 0 sec cpu: 20397.3 % mem: [20668, 21375, 33965] MB
[removing ec, pass 3 ] 100 % elapsed: 0 min 27 sec remaining: 0 min 0 sec cpu: 22962.5 % mem: [20167, 20698, 33965] MB
[removing ec, pass 4 ] 100 % elapsed: 0 min 27 sec remaining: 0 min 0 sec cpu: 23310.6 % mem: [19924, 20170, 33965] MB
[removing ec, pass 5 ] 100 % elapsed: 0 min 28 sec remaining: 0 min 0 sec cpu: 23301.5 % mem: [19860, 19925, 33965] MB
[removing tips, pass 6 ] 100 % elapsed: 0 min 55 sec remaining: 0 min 0 sec cpu: 25142.7 % mem: [19974, 19974, 33965] MB
[removing bulges, pass 6 ] 100 % elapsed: 1 min 36 sec remaining: 0 min 0 sec cpu: 22626.6 % mem: [19948, 20002, 33965] MB
[removing ec, pass 6 ] 100 % elapsed: 0 min 27 sec remaining: 0 min 0 sec cpu: 22708.7 % mem: [19920, 19947, 33965] MB
[removing tips, pass 7 ] 100 % elapsed: 0 min 21 sec remaining: 0 min 0 sec cpu: 24399.3 % mem: [19909, 19921, 33965] MB
[removing bulges, pass 7 ] 100 % elapsed: 1 min 37 sec remaining: 0 min 0 sec cpu: 23080.5 % mem: [19853, 19911, 33965] MB
[removing ec, pass 7 ] 100 % elapsed: 0 min 26 sec remaining: 0 min 0 sec cpu: 23476.6 % mem: [19837, 19848, 33965] MB
[removing tips, pass 8 ] 100 % elapsed: 0 min 16 sec remaining: 0 min 0 sec cpu: 8285.7 % mem: [19837, 19841, 33965] MB
[removing bulges, pass 8 ] 100 % elapsed: 1 min 37 sec remaining: 0 min 0 sec cpu: 23144.2 % mem: [19833, 19837, 33965] MB
[removing ec, pass 8 ] 100 % elapsed: 0 min 26 sec remaining: 0 min 0 sec cpu: 23530.7 % mem: [19827, 19828, 33965] MB
[removing tips, pass 9 ] 100 % elapsed: 0 min 15 sec remaining: 0 min 0 sec cpu: 8222.7 % mem: [19826, 19827, 33965] MB
[removing bulges, pass 9 ] 100 % elapsed: 1 min 36 sec remaining: 0 min 0 sec cpu: 23186.3 % mem: [19831, 19831, 33965] MB
[removing ec, pass 9 ] 100 % elapsed: 0 min 27 sec remaining: 0 min 0 sec cpu: 23680.5 % mem: [19825, 19826, 33965] MB
[removing tips, pass 10 ] 100 % elapsed: 0 min 16 sec remaining: 0 min 0 sec cpu: 8274.7 % mem: [19825, 19826, 33965] MB
[removing bulges, pass 10 ] 100 % elapsed: 1 min 36 sec remaining: 0 min 0 sec cpu: 23208.0 % mem: [19829, 19829, 33965] MB
[removing ec, pass 10 ] 100 % elapsed: 0 min 28 sec remaining: 0 min 0 sec cpu: 23634.8 % mem: [19825, 19825, 33965] MB
[Minia : assembly ] 100 % elapsed: 8 min 6 sec remaining: 0 min 0 sec cpu: 100.2 % mem: [19798, 19798, 33965] MB
[Approximating frequencies of minimizers ] 100 % elapsed: 1 min 17 sec remaining: 0 min 0 sec cpu: 99.8 % mem: [ 28, 28, 28] MB
[DSK: nb solid kmers found : 678092331 ] 201 % elapsed: 88 min 9 sec remaining: 0 min 0 sec cpu: 1630.6 % mem: [7642, 7848, 7871] MB
[Iterating DSK partitions ] 0 % elapsed: 0 min 0 sec remaining: 0 min 0 sec/home/pavel/bin/gatb-minia-pipeline/tools/memused: line 18: 249751 Segmentation fault "$@"

[rchikhi]

Hi Mozart, thanks for reporting it. I'm assuming you cannot share the data either, therefore I'd be curious to see if the problem occurs with a different k-mer combinations. Can you please try the following command line? ./gatb --kmer-sizes 31,51,71 -1 [..] -2 [..] -o [..]

[Mozart1776]

Hi Ryan, Thanks for your response. Actually I would be Ok with me to share the data, if it could be useful for your. Let me know what is the best way to do it for you. All the best! Pavel 2018-08-26 7:41 GMT-04:00 Rayan Chikhi <[email protected]>:

…

Hi Mozart, thanks for reporting it. I'm assuming you cannot share the data either, therefore I'd be curious to see if the problem occurs with a different k-mer combinations. Can you please try the following command line? ./gatb --kmer-sizes 31,51,71 -1 [..] -2 [..] -o [..] — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#6 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AoTCNqdcCoukMZVFE1SiJUbLVIXRhUALks5uUolygaJpZM4SsR9k> .

-- Pavel B. Klimov, Ph.D. University of Michigan, Department of Ecology and Evolutionary Biology, Museum of Zoology 3600 Varsity Dr. #1030, Ann Arbor, Michigan 48108 USA Phone (office): [phone] Fax: [phone] Email (business): [email protected] Web: http://insects.ummz.lsa.umich.edu/ACARI/staff/pklimov Bee mites: http://idtools.org/id/mites/beemites/

[rchikhi]

Hi Pavel,
Very nice! Please email me at [email protected]. Any way to get the data is fine, if you can share it on your end. Otherwise I'll send a shared dropbox link.
Rayan

[adigenova]

HI Rayan,
I have a similar problem with DSK(core dumped), the log is the following:

Minia 3, git commit 4b32fec
setting storage type to hdf5
[Approximating frequencies of minimizers ] 100 % elapsed: 2 min 45 sec remaining: 0 min 0 sec cpu: 99.9 % mem: [ 25, 25, 25] MB
[DSK: Pass 1/1, Step 2: counting kmers ] 73 % elapsed: 145 min 33 sec remaining: 53 min 49 sec cpu: 553.2 % mem: [36580, 36580, 36580] MB /home/adigenova/binaries/gatb-minia-pipeline/tools/memused: line 18: 33019 Segmentation fault (core dumped) "$@"
maximal memory used: 41023 MB
(2018-09-18 19:12:07) Execution of 'minia/minia' failed. Command line:
/home/adigenova/binaries/gatb-minia-pipeline/tools/memused /home/adigenova/binaries/gatb-minia-pipeline/minia/minia -in MMK.list_reads -kmer-size 61 -abundance-min 2 -out MMK_k61 -max-memory 35000

The command line that I used was the following:
~/binaries/gatb-minia-pipeline/gatb -l sequences.txt --nb-cores 20 --max-memory 35000 -o MMK --kmer-sizes 31,61,91 --abundance-mins 2,2,2

and the sequences.txt file contains the following reads:
stLFR.split_read.1.fq.gz
stLFR.split_read.2.fq.gz

the reads are from GIAB:
ftp-trace.ncbi.nlm.nih.gov/giab/ftp/data/NA12878/stLFR/stLFR.split_read.1.fq.gz
ftp-trace.ncbi.nlm.nih.gov/giab/ftp/data/NA12878/stLFR/stLFR.split_read.2.fq.gz

Let me know if you need more information to reproduce the error.
Thank in advance.

The best
Alex

[rchikhi]

Hi Alex,
Thanks for the very detailed bug report and sorry for the answer delay. Can you reproduce this problem on another machine? Unfortunately I cannot, the pipeline finished without crashing on my server using the command line you provided, on the master branch of gatb-pipeline repo.

[adigenova]

Hi Rayan: Thanks for the answer, I was running the command in a cluster and it crushed in different nodes. The node hardware is the following: Intel(R) Xeon(R) CPU E5-2660 v2 @ 2.20GHz, 47Gb of RAM, Lustre filesytem. To me the problem is the memory that the program use, because it ran without problems when the coverage is less than 30X for a human dataset. I mean the BCALM step use less than 35GB of RAM in such datasets. I observed that the same error occur in the following situations: 1.- Coverage is more than 30X for a human genome. 2.- When the length of short reads is 250bp, I got the same error on arabidopsis or human datasets. Of course a solution is to increase the RAM to something like 100Gb, but unfortunately we don’t have such kind of machines. Do you think I have to update to the current master? Thank in advance Best Alex

…

On Oct 29, 2018, at 1:09 PM, Rayan Chikhi ***@***.***> wrote: Hi Alex, Thanks for the very detailed bug report and sorry for the answer delay. Can you reproduce this problem on another machine? Unfortunately I cannot, the pipeline finished without crashing on my server using the command line you provided. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#6 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHexyE68Rjg0e1zr9KcUD41AdWPA6jiUks5upvAVgaJpZM4SsR9k>.

[rchikhi]

After offline discussion with Alex it seems that the problem he reported was related to higher memory usage than what was available on the system.
I'm closing this thread as it's aggregating a bunch of unrelated segfault problems.