cometbft: tests: `make test` fails locally while Github CI checks pass
Bug Report
Setup
Branch main, clean repository:
$ git rev-parse --verify HEAD
2d097e134ff29ee2ad6ed9653b7dc788fe4a0555
$ git status --ignored
On branch main
Your branch is up to date with 'origin/main'.
nothing to commit, working tree clean
Have you tried the latest version: it is the main branch. Did not test on release branches.
ABCI app (name for built-in, URL for self-written if it’s publicly available): not relevant
Environment:
- OS (e.g. from /etc/os-release): Darwin 22.6.0 Darwin Kernel Version 22.6.0: Fri Sep 15 13:41:28 PDT 2023; root:xnu-8796.141.3.700.8~1/RELEASE_ARM64_T6000 arm64
- Install tools:
go version go1.21.3 darwin/arm64
Config
Standard.
How to reproduce it
- Download the repository
- Run
make test - Go grab (or make) a coffee
What happened?
Some tests fail, after very very long 18:47.13 minutes.
What did you expect to happen?
The tests succeed. Maybe the execution to be more expedite.
Logs
make test &> output.txt 1083.17s user 326.12s system 125% cpu 18:47.13 total
Output: https://gist.github.com/cason/036e2fe1738157f2a452d9bb588ff051
dump_consensus_state output
Irrelevant.
Anything else we need to know
Checks performed by Github CI pass, including unit tests. Refer, for instance, to a recent pull request, like https://github.com/cometbft/cometbft/pull/2250.
About this issue
- Original URL
- State: open
- Created 3 years ago
- Reactions: 1
- Comments: 16 (16 by maintainers)
Commits related to this issue
- chore(tests): enable and cleanup integration tests Refs #2260 — committed to cometbft/cometbft by melekes 2 years ago
- chore(tests): cleanup integration tests (#4088) Refs #2260 --------- Co-authored-by: Andy Nogueira <me@andynogueira.dev> Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com... — committed to cometbft/cometbft by melekes 2 years ago
It does not fail in CI because CI sets the CI environment variable, which causes the test to be skipped. To succeed locally, to
CI=1 make testor similar.Developers reading the tests and seeing
checkTxsInOrderbeing used all around might think FIFO is guaranteed. I would keep the test here, as a reminder that this is not true.My previous suggestion doesn’t improve a lot:
Moreover, tests still fail.
Apparently, the CI job uses the Makefile target
test-group. Which usesgo test -mod=readonly -timeout=15m -race -coverprofile=$(BUILDDIR)/$*.profile.outas testing command. Maybemake testshould use the same command.I created the somehow related PR #2263 just before seeing this issue 🙂