nic-6443 commented on PR #13251: URL: https://github.com/apache/apisix/pull/13251#issuecomment-4269851912
CI ping — failures on the two reruns both look like pre-existing flakes, not something this PR caused. Wanted to document the evidence before triggering another rerun. ### `build (linux_openresty, t/admin t/cli ...)` — `t/admin/plugins-reload.t` TEST 1 + TEST 2 Same test failed on master run [24491019864](https://github.com/apache/apisix/actions/runs/24491019864) (2026-04-16) with the identical symptom: - TEST 1: `grep_error_log_out` got `''`, expected `sync local conf to etcd` - TEST 2: `reload plugins on node before reload` line missing This PR doesn't touch `apisix/admin/`, `apisix/core/`, or `apisix/plugin.lua`. Diff is Lua files under `apisix/plugins/ai-proxy*` + docs + one new `.t` file. Root cause looks like a race between admin's startup `sync_local_conf_to_etcd(true)` timer and the test's assertions — when `/plugins` is synced fast enough the log/filter fires before the test's `grep_error_log` window opens, and the test sees nothing. Fix belongs in the test (make it deterministic) but out of scope here. ### `build (linux_apisix_current_luarocks)` — `t/cli/test_etcd_sync_event_handle.sh` ``` + etcdctl --user=root:apache-api6-sync put /apisix/routes/1 ... Error: context deadline exceeded + clean_up ``` The script does `make ci-env-stop && make ci-env-up` and then immediately issues an `etcdctl put` without waiting for the fresh etcd container to be ready. `etcdctl` times out and `set -e` aborts. Pre-existing infra race in the test script itself. Triggering another rerun. Happy to open a separate issue to track either flake if that's useful. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
