* fix: use `exit-nopipe` with consistent `shell` usage
Temporarily disabled the `set -e` option around the docker logs command to handle the broken pipe error gracefully.
Handle more complex scenarios in our `Result of ${{ inputs.test_id }} test` job
* fix: Use single quotes for the outer command
* fix: use same approach for CD
* test: check launch failure logs
* fix: revert CD changes
* fix: do not try to increase the disk size and wait mounting
* fix: increase GB a bit more
* fix: do not fail on pipe failure
* fix: use plain `tee /dev/stderr`
If this does not work try `(tee … || true)`
* fix: `tee` not stoping on cd config tests
* fix: match logic with GCP tests
* fix(cd): handle pipe and other errors correctly
* try `tee --output-error=exit-nopipe`
* fix: TRAP without pipefail
* test: pipefail with exit and trap
* fix: use a subshell
* fix(ci): wait for mounting and show system logs if fail
* fix(ci): GCP is not always mounting disks in the same order
* fix: use `grep` instead of `awk`
* fix: typo
* fix: use simpler `grep` command
* fix: do not sleep if not require
* chore: reduce diff
|
||
|---|---|---|
| .. | ||
| ISSUE_TEMPLATE | ||
| PULL_REQUEST_TEMPLATE | ||
| workflows | ||
| CODEOWNERS | ||
| dependabot.yml | ||
| mergify.yml | ||
| pull_request_template.md | ||
| release-drafter.yml | ||
| zbot.yml | ||