chore: disable failpoint test in simulation integration test (#13962)

Makefile.toml

@@ -965,7 +965,6 @@ set -e
 
 cargo nextest run \
   --config "target.'cfg(all())'.rustflags = ['--cfg=madsim']" \
-  --features fail/failpoints \
   -p risingwave_simulation \
   "$@"
 """
@@ -981,7 +980,6 @@ set -e
 
 cargo nextest archive \
   --config "target.'cfg(all())'.rustflags = ['--cfg=madsim']" \
-  --features fail/failpoints \
   -p risingwave_simulation \
   --archive-file simulation-it-test.tar.zst \
   "$@"

ci/workflows/pull-request.yml

@@ -63,8 +63,6 @@ steps:
           run: rw-build-env
           config: ci/docker-compose.yml
           mount-buildkite-agent: true
-    # FIXME: decrease this when we found what's causing the build to be slow.
-    timeout_in_minutes: 20
     retry: *auto-retry
 
   - label: "docslt"

src/tests/simulation/tests/integration_tests/recovery/event_log.rs

@@ -128,6 +128,7 @@ async fn test_create_fail(
 
 /// Tests event log can record info of stream job creation failure, for CREATE TABLE/MV/INDEX/SINK.
 #[tokio::test]
+#[ignore]
 async fn failpoint_limited_test_create_stream_job_fail() -> Result<()> {
     let mut cluster = Cluster::start(cluster_config()).await.unwrap();
     let mut session = cluster.start_session();
@@ -224,6 +225,7 @@ async fn failpoint_limited_test_create_stream_job_fail() -> Result<()> {
 /// Theoretically errors either reported by compute nodes, or chosen by meta node, may not be the root cause.
 /// But during my tests with MADSIM_TEST_SEED from 1..1000, this test always succeeded.
 #[tokio::test]
+#[ignore]
 async fn failpoint_limited_test_collect_barrier_failure() -> Result<()> {
     let mut cluster = Cluster::start(cluster_config()).await.unwrap();
     let mut session = cluster.start_session();