Add error handling for sgx ci #4222
Conversation
lum1n0us
requested review from
loganek,
no1wudi,
TianlongLiang,
wenyongh,
xujuntwt95329 and
yamt
as code owners
lum1n0us
force-pushed
the
fix/capture_ci_143_code
branch
from
659d11d to
a2c5637
Compare
> Process completed with exit code 143. It is a known issue with GitHub-hosted runners. Usually, increasing the swap file can help avoid it. However, sometimes error 143 still occurs. To prevent confusion, let's capture error 143 and allow the CI to pass.
lum1n0us
force-pushed
the
fix/capture_ci_143_code
branch
from
a2c5637 to
ec65023
Compare
| if [ $exitcode -eq 143 ]; then | ||
| echo "$exitcode is a known GitHub-hosted runner issue" | ||
| echo "::notice::Bypass an error with code 143 in SGX CI" | ||
| exit 0 |
There was a problem hiding this comment.
doesn't this risk us to merge broken changes?
There was a problem hiding this comment.
I believe it will not introduce any other unexpected breaking changes.
- Only bypass exit code 143. Other non-zero codes will still be captured.
- Spec_test scripts, all.py and runtest.py, will clamp return codes, and 143 is not one of them.
- 143 is not an errno in errno.h
There was a problem hiding this comment.
well, my concern is:
- someone submits a PR with broken changes
- the ci happened to pass because of this workaround
- reviewers don't notice it and merge
There was a problem hiding this comment.
Based on my observations, if there is an error from the spec-test itself, 143 won't be generated. I believe the 143 code will not cover other errors. Those accidental 143 errors are quite annoying, so I want to relieve our daily work. If directly bypassing is too aggressive, how about automatically rerunning if it fails?
There was a problem hiding this comment.
Those accidental 143 errors are quite annoying, so I want to relieve our daily work.
i agree.
If directly bypassing is too aggressive, how about automatically rerunning if it fails?
it seems nicer.
do you have a reference to a bug ticket or something? |
actions/runner-images#6680, there is also an approach we can try: adding |
large runners require some extra money, don't they? |
It is a known issue with GitHub-hosted runners. Usually, increasing the swap file can help avoid it. However, sometimes error 143 still occurs. To prevent confusion, let's capture error 143 and allow the CI to pass.