trRosetta2 icon indicating copy to clipboard operation
trRosetta2 copied to clipboard

HHSearch: 7701 Segmentation fault (core dumped)

Open cclough opened this issue 3 years ago • 12 comments

I'm getting an error on the hhsearch step:

Running hhsearch ./run_pipeline.sh: line 49: 7701 Segmentation fault (core dumped) $HH -i $WDIR/t000_.msa0.ss2.a3m -o $WDIR/t000_.hhr -v 0 > $WDIR/log/hhsearch.stdout 2> $WDIR/log/hhsearch.stderr

There's no output in hhsearch.stderr

It ran fine for T1078, but not for my 464aa protein.

Running the HHsearch step in isolation shows the following:

- 04:14:18.171 INFO: Searching 136710 database HHMs without prefiltering

- 04:14:18.221 INFO: Iteration 1

- 04:14:18.332 INFO: Scoring 136710 HMMs using HMM-HMM Viterbi alignment

- 04:14:18.614 INFO: Alternative alignment: 0

- 04:14:33.097 INFO: 136710 alignments done

- 04:14:33.333 INFO: Alternative alignment: 1

- 04:14:33.926 INFO: 3511 alignments done

- 04:14:33.929 INFO: Alternative alignment: 2

- 04:14:33.967 INFO: 188 alignments done

- 04:14:33.967 INFO: Alternative alignment: 3

- 04:14:33.982 INFO: 2 alignments done

Segmentation fault (core dumped)

I've tried things suggested elsewhere via google e.g. sudo sysctl vm.overcommit_memory=1, but no luck.

The protein returns a very large number of sequences from HHBlits. Initially it crashed there and I had to limit the MSA size with hhblits -maxseq 500

I have HHsearch v3.3.0

16 CPUs, 96GB memory. I've tried with more and less.

cclough avatar Jun 21 '21 13:06 cclough

Same here: the T1078 example ran fine but I also got the same problem for my sequence of only 200 aa. Have you @cclough figured out yet? Pls post the solutions if you know. Thanks!

jennifermbui avatar Jun 22 '21 20:06 jennifermbui

@jennifermbui no still blocked on this

cclough avatar Jun 23 '21 11:06 cclough

@cclough, I actual installed HH-suite3 on my linux desktop rather using the command under miniconda. I'm sure that Sergey @lyskov and Ivan @gjoni would know why it broke. But, the name of the game is to produce t000_.hhr.

jennifermbui avatar Jun 30 '21 23:06 jennifermbui

I found the same issue and got in contact with the developers of hh-suite. After tracing back the part of the code that was returning the segmentation fault, it seems PDB100 database does not contain an _a3m database which is required by the last hhsearch version. Is it possible to make this database available with the PDB100 database?

ajrferrari avatar Jul 05 '21 12:07 ajrferrari

They added a flag which ignores the a3m missing file. Update your from source compiled version and add -premerge 0.

$HH -i $WDIR/t000_.msa0.ss2.a3m -o $WDIR/t000_.hhr -premerge 0

ajrferrari avatar Jul 05 '21 13:07 ajrferrari

Thanks for following up with hh-suite developers, @ajrferrari! @lyskov, I wonder if you would add to this flag to run_pipeline.sh. Also, I'm not sure why PDB100 does not contain an _a3m database. @aleaverfay @lyskov, Is it possible to make this database available with the PDB100 database? Thanks! Jen

jennifermbui avatar Jul 05 '21 17:07 jennifermbui

Thanks for following up with hh-suite developers, @ajrferrari! @lyskov, I wonder if you would add to this flag to run_pipeline.sh. Also, I'm not sure why PDB100 does not contain an _a3m database. @aleaverfay @lyskov, Is it possible to make this database available with the PDB100 database? Thanks! Jen

re flags addition and database alteration: tagging original authors @gjoni @minkbaek

lyskov avatar Jul 06 '21 21:07 lyskov

The use of -premerge 0 also require to use hhsearch last version available at https://github.com/soedinglab/hh-suite

FrBonnardel avatar Jul 08 '21 11:07 FrBonnardel

You can use the v3.3.0 if you compile from the github source. The premerge option was made available in the last couple days. Please pull the updated version.

On Thu, Jul 8, 2021 at 6:36 AM FrBonnardel @.***> wrote:

The use of -premerge 0 also require to use hhsearch last version

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/RosettaCommons/trRosetta2/issues/7#issuecomment-876365694, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFZL64PVNQ4TPWM342AKVKTTWWETHANCNFSM47BTNP2A .

-- Állan

ajrferrari avatar Jul 08 '21 11:07 ajrferrari

it still not working with me It shows an error in : " Running hhsearch ./run_pipeline.sh: line 57: 131266 Segmentation fault (core dumped) $HH -i $WDIR/t000_.msa0.ss2.a3m -o $WDIR/t000_.hhr -premerge 0 > $WDIR/log/hhsearch.stdout 2> $WDIR/log/hhsearch.stderr "

With the installed version HHsearch 3.3.0 Although it was worked well and outputs a result in an example T1078 my PC is core i7 and RAM 8G

menanagy01 avatar Jul 28 '21 09:07 menanagy01

it still not working with me It shows an error in : " Running hhsearch ./run_pipeline.sh: line 57: 131266 Segmentation fault (core dumped) $HH -i $WDIR/t000_.msa0.ss2.a3m -o $WDIR/t000_.hhr -premerge 0 > $WDIR/log/hhsearch.stdout 2> $WDIR/log/hhsearch.stderr "

With the installed version HHsearch 3.3.0 Although it was worked well and outputs a result in an example T1078 my PC is core i7 and RAM 8G

i have the same problam ,can you tell me your solution?

IL-6Li avatar Dec 21 '23 05:12 IL-6Li

Hi, it can be due to diverse issues including lack of memory, disk space, wrong version and corrupted patterns file.CheersLe 21 déc. 2023 06:06, giaoli @.***> a écrit :

it still not working with me It shows an error in : " Running hhsearch ./run_pipeline.sh: line 57: 131266 Segmentation fault (core dumped) $HH -i $WDIR/t000_.msa0.ss2.a3m -o $WDIR/t000_.hhr -premerge 0 > $WDIR/log/hhsearch.stdout 2> $WDIR/log/hhsearch.stderr " With the installed version HHsearch 3.3.0 Although it was worked well and outputs a result in an example T1078 my PC is core i7 and RAM 8G

i have the same problam ,can you tell me your solution?

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>

FrBonnardel avatar Dec 21 '23 07:12 FrBonnardel