app_test_seqan_tcoffee fails on ppc64le #2365

paulmenzel · 2018-12-04T14:47:24Z

$ make test
[…]
        Start 379: build_dox
379/434 Test #379: build_dox .........................................................................***Failed    0.15 sec
[…]
421/434 Test #421: app_test_seqan_tcoffee ............................................................***Failed    1.52 sec
        Start 422: ctd_test_seqan_tcoffee
[…]
99% tests passed, 2 tests failed out of 434

Total Test time (real) = 342.71 sec

The following tests FAILED:
        379 - build_dox (Failed)
        421 - app_test_seqan_tcoffee (Failed)
Errors while running CTest
make: *** [Makefile:109: test] Error 8

The text was updated successfully, but these errors were encountered:

paulmenzel · 2018-12-04T14:47:50Z

$ git log --oneline -1
294edb288 Merge pull request #2361 from joergi-w/fix/empty_graph_iterator

rrahn · 2018-12-05T10:28:34Z

Thanks for reporting.
Can you check if this still occurs when #2364 is merged.

paulmenzel · 2018-12-05T11:12:57Z

It still does occur.

$ git log -1
commit 4b5169376808dbc688a3fadd754bd0c0e437af55 (HEAD -> master, origin/master, origin/HEAD)
Author: rrahn <rene.rahn@fu-berlin.de>
Date:   Fri Nov 30 11:40:49 2018 +0100

    [FIX] Use default construction to initilaize data members.
$ make test
[…]
99% tests passed, 3 tests failed out of 434

Total Test time (real) = 391.11 sec

The following tests FAILED:
	264 - test_demo_tutorial_journaled_set_solution_online_search_finder (Failed)
	379 - build_dox (Failed)
	421 - app_test_seqan_tcoffee (Failed)
Errors while running CTest
make: *** [Makefile:109: test] Error 8

paulmenzel · 2018-12-05T11:14:49Z

Where can I find the logs of the failed tests?

rrahn · 2018-12-05T11:15:57Z

Can you rerun with ctest option --output-on-failure. Then we see the failing tests and can try to investigate this.

paulmenzel · 2018-12-05T11:39:21Z

For failing build_dox I created #2366.

[…]
        Start 421: app_test_seqan_tcoffee
421/434 Test #421: app_test_seqan_tcoffee ............................................................***Failed    1.50 sec
Comparing /dev/shm/seqan/apps/seqan_tcoffee/tests/1aab.mlocal.fasta against /tmp/tmpe9mbl8/1aab.mlocal.fasta
---

+++

@@ -1,8 +1,8 @@

 >hmgl_trybr 
-KKDSNAPKRAMTSFMFFSSDFRS----KHSDLS-IVEMSKAAGAAWKELGPEERKVYEEM
+KKDSNAPKRAMTSFMFFSSDFRS-KHSD---LS-IVEMSKAAGAAWKELGPEERKVYEEM
 AEKDKERYKREM----------------
 >hmgt_mouse 
------KPKRPRSAYNIYVSESFQEAKDDSAQGK-----LKLVNEAWKNLSPEEKQAYIQL
+-----KPKRPRSAYNIYVSESFQEAKDD----S-AQGKLKLVNEAWKNLSPEEKQAYIQL
 AKDDRIRYDNEMKSWEEQMAE-------
 >hmgb_chite 
 ---ADKPKRPLSAYMLWLNSARESIKRENPDFK-VTEVAKKGGELWRGL--KDKSEWEAK
DELETING TEMP DIR /tmp/tmpe9mbl8
Executing test for seqan_tcoffee
================================

seqan_tcoffee -s /dev/shm/seqan/apps/seqan_tcoffee/tests/1aab.fa -o /tmp/tmpe9mbl8/1aab.fasta OK
seqan_tcoffee -s /dev/shm/seqan/apps/seqan_tcoffee/tests/1ad2.fa -o /tmp/tmpe9mbl8/1ad2.fasta OK
seqan_tcoffee -s /dev/shm/seqan/apps/seqan_tcoffee/tests/2trx.fa -o /tmp/tmpe9mbl8/2trx.fasta OK
seqan_tcoffee -a protein -s /dev/shm/seqan/apps/seqan_tcoffee/tests/1aab.fa -o /tmp/tmpe9mbl8/1aab.protein.fasta OK
seqan_tcoffee -a protein -s /dev/shm/seqan/apps/seqan_tcoffee/tests/1ad2.fa -o /tmp/tmpe9mbl8/1ad2.protein.fasta OK
seqan_tcoffee -a protein -s /dev/shm/seqan/apps/seqan_tcoffee/tests/2trx.fa -o /tmp/tmpe9mbl8/2trx.protein.fasta OK
seqan_tcoffee -m global -s /dev/shm/seqan/apps/seqan_tcoffee/tests/1aab.fa -o /tmp/tmpe9mbl8/1aab.mglobal.fasta OK
seqan_tcoffee -m local -s /dev/shm/seqan/apps/seqan_tcoffee/tests/1aab.fa -o /tmp/tmpe9mbl8/1aab.mlocal.fasta FAILED
seqan_tcoffee -m overlap -s /dev/shm/seqan/apps/seqan_tcoffee/tests/1aab.fa -o /tmp/tmpe9mbl8/1aab.moverlap.fasta OK
[…]
seqan_tcoffee -a rna -s /dev/shm/seqan/apps/seqan_tcoffee/tests/adeno4-rna.fa -o /tmp/tmpe9mbl8/adeno4-rna.fasta OK
==============================
     total tests: 66
    failed tests: 1
successful tests: 65
==============================

rrahn · 2018-12-05T11:50:50Z

@joergi-w can you please try to investigate this problem. @paulmenzel is there any possibilty we can get remote access to a ppc64le in case we can't reproduce it on our platforms?

paulmenzel · 2018-12-05T11:56:25Z

On 12/05/18 12:50, rrahn wrote: @paulmenzel is there any possibilty we can get remote access to a ppc64le in case we can't reproduce it on our platforms?

Yes, we could figure something out and set up a guest account.

rrahn · 2018-12-05T12:38:07Z

Awsome

joergi-w · 2018-12-05T14:46:33Z

Sure, I'll look into this!
I actually have an account in the MPI-MG; I just need the host name of a ppc64le machine where I can log in to perform the builds and tests.

paulmenzel · 2018-12-05T14:53:31Z

On 12/05/18 15:46, Jörg Winkler wrote: I actually have an account in the MPI-MG; I just need the host name of a ppc64le machine where I can log in to perform the builds and tests.

I’ll send you a private message.

joergi-w · 2018-12-06T15:19:25Z

Very strange: In Release build mode all tests pass successfully and in Debug mode I see 5 different tests fail, where app_test_seqan_tcoffee is not contained.

99% tests passed, 5 tests failed out of 434

Total Test time (real) = 4577.06 sec

The following tests FAILED:
        264 - test_demo_tutorial_journaled_set_solution_online_search_finder (Failed)
        403 - app_test_ngs_roi (Failed)
        410 - app_test_razers (Failed)
        411 - app_test_razers3 (Failed)
        412 - app_test_razers3_sequential (Failed)

I'll check out an earlier version to test whether fix/empty_graph_iterator (#2361) introduced the problem. However with the result shown above it doesn't look like an issue specific to T-Coffee.

joergi-w · 2018-12-10T12:01:26Z

Also in version d7b7124 multiple tests fail (in Debug mode only).

@paulmenzel Which compiler flags do you use? I still cannot reproduce the tcoffee failure, but spotted some other tests that do not succeed on ppc64le. Do you remember a time point/ commit in the past, where all tests were successful for you?

paulmenzel · 2018-12-10T12:06:03Z

On 12/10/18 13:01, Jörg Winkler wrote: Also in version d7b7124 multiple tests fail (in Debug mode only).

Thank you for looking into that.

@paulmenzel Which compiler flags do you use? I still cannot reproduce the _tcoffee_ failure, but spotted some other tests that do not succeed on _ppc64le_.

I am using the defaults. I built in `/dev/shm` and run the commands below. $ mkdir build $ cd build $ cmake ../ $ make -j120 && make test

Do you remember a time point/ commit in the past, where all tests were successful for you?

No, I just started to build and test that now, and have no data from the past.

joergi-w · 2018-12-11T12:41:01Z

Thank you for the information! Your binary in /dev/shm/... indeed fails the test.

However, I use the same commands in my home directory on your machine and the tcoffee test still works for me (and all the other tests, too). I suggest to try a new build from the latest develop branch in a different directory. Sorry that I currently cannot investigate more.

paulmenzel · 2018-12-12T15:09:24Z

Didn’t you write, that tests failed for you too?

Anyway, building in my home directory, tests still fail, but unfortunately not reproducibly.

$ make test
[…]
The following tests FAILED:
        264 - test_demo_tutorial_journaled_set_solution_online_search_finder (Failed)
        421 - app_test_seqan_tcoffee (Failed)

Running ctest . --output-on-failure directly, I only got the output below.

        Start 421: app_test_seqan_tcoffee
421/434 Test #421: app_test_seqan_tcoffee ............................................................***Failed    1.49 sec
Comparing /home/pmenzel/seqan/apps/seqan_tcoffee/tests/1aab.mlocal.fasta against /tmp/tmp8ZnHVf/1aab.mlocal.fasta
---

+++

@@ -1,8 +1,8 @@

 >hmgl_trybr
-KKDSNAPKRAMTSFMFFSSDFRS----KHSDLS-IVEMSKAAGAAWKELGPEERKVYEEM
+KKDSNAPKRAMTSFMFFSSDFRS-KHSD---LS-IVEMSKAAGAAWKELGPEERKVYEEM
 AEKDKERYKREM----------------
 >hmgt_mouse
------KPKRPRSAYNIYVSESFQEAKDDSAQGK-----LKLVNEAWKNLSPEEKQAYIQL
+-----KPKRPRSAYNIYVSESFQEAKDD----S-AQGKLKLVNEAWKNLSPEEKQAYIQL
 AKDDRIRYDNEMKSWEEQMAE-------
 >hmgb_chite
 ---ADKPKRPLSAYMLWLNSARESIKRENPDFK-VTEVAKKGGELWRGL--KDKSEWEAK
DELETING TEMP DIR /tmp/tmp8ZnHVf
Executing test for seqan_tcoffee
================================
[…]
seqan_tcoffee -m local -s /home/pmenzel/seqan/apps/seqan_tcoffee/tests/1aab.fa -o /tmp/tmp8ZnHVf/1aab.mlocal.fasta FAILED
[…]

rrahn mentioned this issue Dec 5, 2018

[FIX] Fixes the ctd tests on travis. #2364

Merged

rrahn assigned joergi-w Dec 5, 2018

rrahn added the bug label Dec 5, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

app_test_seqan_tcoffee fails on ppc64le #2365

app_test_seqan_tcoffee fails on ppc64le #2365

paulmenzel commented Dec 4, 2018

paulmenzel commented Dec 4, 2018

rrahn commented Dec 5, 2018

paulmenzel commented Dec 5, 2018 •

edited

paulmenzel commented Dec 5, 2018

rrahn commented Dec 5, 2018

paulmenzel commented Dec 5, 2018

rrahn commented Dec 5, 2018

paulmenzel commented Dec 5, 2018 via email

rrahn commented Dec 5, 2018

joergi-w commented Dec 5, 2018

paulmenzel commented Dec 5, 2018 via email

joergi-w commented Dec 6, 2018

joergi-w commented Dec 10, 2018

paulmenzel commented Dec 10, 2018 via email

joergi-w commented Dec 11, 2018

paulmenzel commented Dec 12, 2018

app_test_seqan_tcoffee fails on ppc64le #2365

app_test_seqan_tcoffee fails on ppc64le #2365

Comments

paulmenzel commented Dec 4, 2018

paulmenzel commented Dec 4, 2018

rrahn commented Dec 5, 2018

paulmenzel commented Dec 5, 2018 • edited

paulmenzel commented Dec 5, 2018

rrahn commented Dec 5, 2018

paulmenzel commented Dec 5, 2018

rrahn commented Dec 5, 2018

paulmenzel commented Dec 5, 2018 via email

rrahn commented Dec 5, 2018

joergi-w commented Dec 5, 2018

paulmenzel commented Dec 5, 2018 via email

joergi-w commented Dec 6, 2018

joergi-w commented Dec 10, 2018

paulmenzel commented Dec 10, 2018 via email

joergi-w commented Dec 11, 2018

paulmenzel commented Dec 12, 2018

paulmenzel commented Dec 5, 2018 •

edited