It turns out there is no test currently to check that the RPC server does not time out under load. With "load" I mean a flood of trivial payloads. That is, the only work needed is JSON encoding and decoding of (let's say) a block of data of 2 MB or so. This may take a few milliseconds, but should never take more than a few seconds.
So add a test for this.
<!--e57a25ab6845829454e8d69fc972939a-->
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.
<!--021abf342d371248e50ceaed478a90ca-->
See the guideline for information on the review process.
| Type | Reviewers |
|---|---|
| Concept ACK | enirox001 |
If your review is incorrectly listed, please copy-paste <code><!--meta-tag:bot-skip--></code> into the comment that the bot should ignore.
<!--174a7506f384e20aa4161008e828411d-->
Reviewers, this pull request conflicts with the following ones:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.
<!--5faf32d7da4f0f540f40219e4f7537a3-->
Draft for now, because the test fails :(
Edit: It fails in CI, but it passes locally for me:
./bld-cmake/test/functional/test_runner.py -j 16 $( printf 'rpc_echo_payload.py %.0s' {1..2000} )
Timed out after 40 minutes in https://github.com/bitcoin/bitcoin/actions/runs/23606118115/job/68749250793?pr=34927#step:11:3195:
2026-03-26T16:46:08.3589405Z Remaining jobs: [interface_ipc_mining.py, rpc_echo_payload.py]
2026-03-26T16:47:46.9801923Z ......................................................................................................................................................................................................
2026-03-26T16:47:47.4805641Z
2026-03-26T16:47:47.4806053Z Remaining jobs: [rpc_echo_payload.py]
2026-03-26T17:27:36.9551875Z ##[error]The operation was canceled.
cc @dergoegge Can you copy-paste the Python test into Antithesis to see if it fails there?
Can this be tested with 120 commits or so?
Can this be tested with 120 commits or so?
Sorry, turns out I was wrong, and I spun up a new connection while disabling Nagle's. Obviously spinning up a new connection is already known to work around the issue and Nagle's didn't seem to affect it.
<!--9cd9c72976c961c55c7acef8f6ba82cd-->
Re-using the same rpc connection on multiple threads is obviously
unsafe, so this helper can be used to create one connection per thread.
This refactor does not change any behavior and can be reviewed with the
git options:
--color-moved=dimmed-zebra --color-moved-ws=ignore-all-space
Concept ACK
Turning into draft while CI is red, but the code should be correct, reviewable, and mergeable after libevent is nuked.