CI: Increase timeout for test-install test suite.

[csadorf]

The current timeout of 35min is regularly exceeded resulting in test failures. This change increases it to 55min.

[codecov]

Codecov Report

Merging #4697 (2b438ce) into develop (48fa475) will increase coverage by 0.01%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##           develop    #4697      +/-   ##
===========================================
+ Coverage    79.62%   79.63%   +0.01%     
===========================================
  Files          484      484              
  Lines        35831    35831              
===========================================
+ Hits         28527    28529       +2     
+ Misses        7304     7302       -2     

Flag	Coverage Δ
django	73.87% <ø> (ø)
sqlalchemy	73.06% <ø> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
aiida/engine/daemon/client.py	75.65% <0.00%> (+1.04%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 48fa475...2b438ce. Read the comment docs.

[csadorf]

There appears to be some issue with the pre-commit upstream. I'll just rebase and rerun this later today.

[chrisjsewell]

Hopefully this helps, but this is not really the underlying cause of the issue. It is because we have too many test jobs, which ends up causing some to go very slow (or even hang)

[csadorf]

Hopefully this helps, but this is not really the underlying cause of the issue. It is because we have too many test jobs, which ends up causing some to go very slow (or even hang)

This is primarily to address the failure of scheduled nightly runs which should not really be affected by our general workload.

[csadorf]

As a related note, we could host our own runner which might alleviate the situation.

[chrisjsewell]

which should not really be affected by our general workload.

Oh it is definitely to do with the work load; the pytest suite only took < 12 minutes to run for all other jobs in the matrix, but the failing job got killed at 31 minutes: https://github.com/aiidateam/aiida-core/runs/1803637702?check_suite_focus=true. I have had this happen numerous times in the last few weeks.

No problem to increase the timeout, but I would also suggest for this workflow, you can now drop the python 3.6 jobs, inline with #4691

As a related note, we could host our own runner which might alleviate the situation.

I was looking into that for replacing Jenkins, but it has a number of security implications, and in the documentation it basically suggests to only use your own runner on private repositories

[csadorf]

which should not really be affected by our general workload.

Oh it is definitely to do with the work load; the pytest suite only took < 12 minutes to run for all other jobs in the matrix, but the failing job got killed at 31 minutes: https://github.com/aiidateam/aiida-core/runs/1803637702?check_suite_focus=true. I have had this happen numerous times in the last few weeks.

What I mean is that the runners reserved for this repository should be exclusively used for the scheduled test-install workflow in the middle night and it shouldn't be affected by any other CI work.

No problem to increase the timeout, but I would also suggest for this workflow, you can now drop the python 3.6 jobs, inline with #4691

I am planning on tackling that as part of the upcoming coding week.

As a related note, we could host our own runner which might alleviate the situation.

I was looking into that for replacing Jenkins, but it has a number of security implications, and in the documentation it basically suggests to only use your own runner on private repositories

Oh wow, that's really bad. I hope that they will work on that.

[chrisjsewell]

let me know when you have fixed the pre-commit 😄

[chrisjsewell]

aiida/tools/graph/graph_traversers.py:212: error: Incompatible types in assignment (expression has type "float", variable has type "Optional[int]") [assignment]
Found 1 error in 1 file (checked 55 source files)

Started seeing this in other PRs as well (cc @ltalirz), possible dependency version change?

[chrisjsewell]

It appears numpy.inf has changed from being identified as an int to a float 🤷

[chrisjsewell]

Fixing in #4700

[ltalirz]

yes, numpy 1.20 has been released https://numpy.org/doc/1.20/release/1.20.0-notes.html

already broke pymatgen ;-)

[chrisjsewell]

already broke pymatgen ;-)

No surprise there 😆

[csadorf]

I believe this problem is actually related to #4731 which I would like to prioritize to resolve.