Skip to content

Commit d734f53

Browse files
committed
capture tensorflow EESSI test output
1 parent 854e491 commit d734f53

File tree

1 file changed

+18
-7
lines changed

1 file changed

+18
-7
lines changed

ansible/ci/check_eessi.yml

Lines changed: 18 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -17,20 +17,31 @@
1717
repo: "https://github.com/eessi/eessi-demo.git"
1818
dest: "{{ eessi_test_rootdir }}/eessi-demo"
1919

20-
- name: Run test job
21-
ansible.builtin.shell:
22-
cmd: |
20+
- name: Create batch script
21+
copy:
22+
dest: "{{ eessi_test_rootdir }}/eessi-demo/TensorFlow/job.sh"
23+
content: |
24+
#!/usr/bin/env bash
25+
#SBATCH --output=%x.out
26+
#SBATCH --error=%x.out
2327
source /cvmfs/pilot.eessi-hpc.org/latest/init/bash
2428
srun ./run.sh
29+
30+
- name: Run test job
31+
ansible.builtin.shell:
32+
cmd: sbatch --wait --job-name=tensorflow job.sh
2533
chdir: "{{ eessi_test_rootdir }}/eessi-demo/TensorFlow"
26-
executable: /bin/bash
34+
# executable: /bin/bash
2735
register: job_output
28-
async: "{{ 10 * 60 }}" # wait for up to 10 minutes
29-
poll: 15 # check every 15 seconds
36+
37+
- name: Retrieve output
38+
slurp:
39+
src: "{{ eessi_test_rootdir }}/eessi-demo/TensorFlow/tensorflow.out"
40+
register: _tensorflow_out
3041

3142
- name: Fail if job output contains error
3243
fail:
3344
# Note: Job prints live progress bar to terminal, so use regex filter to remove this from stdout
3445
msg: "Test job using EESSI modules failed. Job output was: {{ job_output.stdout | regex_replace('\b', '') }}"
35-
when: '"Epoch 5/5" not in job_output.stdout'
46+
when: '"Epoch 5/5" not in _tensorflow_out.content | b64decode'
3647

0 commit comments

Comments
 (0)