Use buffer tag to avoid hash conflicts in last written LSN cache

.github/workflows/build_and_test.yml

@@ -11,9 +11,8 @@ defaults:
     shell: bash -ex {0}
 
 concurrency:
-  # Allow only one workflow per any non-`main` branch.
-  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.ref == 'refs/heads/main' && github.sha || 'anysha' }}
-  cancel-in-progress: true
+   group: ${{ github.workflow }}-${{ github.ref }}
+   cancel-in-progress: true
 
 env:
   RUST_BACKTRACE: 1

.github/workflows/codestyle.yml

@@ -11,9 +11,8 @@ defaults:
     shell: bash -ex {0}
 
 concurrency:
-  # Allow only one workflow per any non-`main` branch.
-  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.ref == 'refs/heads/main' && github.sha || 'anysha' }}
-  cancel-in-progress: true
+   group: ${{ github.workflow }}-${{ github.ref }}
+   cancel-in-progress: true
 
 env:
   RUST_BACKTRACE: 1

.github/workflows/pg_clients.yml

@@ -13,9 +13,8 @@ on:
   workflow_dispatch:
 
 concurrency:
-  # Allow only one workflow per any non-`main` branch.
-  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.ref == 'refs/heads/main' && github.sha || 'anysha' }}
-  cancel-in-progress: true
+   group: ${{ github.workflow }}-${{ github.ref }}
+   cancel-in-progress: true
 
 jobs:
   test-postgres-client-libs:

pageserver/src/page_service.rs

@@ -716,7 +716,7 @@ impl PageServerHandler {
         let _enter = info_span!("get_nblocks", rel = %req.rel, req_lsn = %req.lsn).entered();
         let latest_gc_cutoff_lsn = timeline.tline.get_latest_gc_cutoff_lsn();
         let lsn = Self::wait_or_get_last_lsn(timeline, req.lsn, req.latest, &latest_gc_cutoff_lsn)?;
-
+		assert!(req.latest);
         let n_blocks = timeline.get_rel_size(req.rel, lsn)?;
 
         Ok(PagestreamBeMessage::Nblocks(PagestreamNblocksResponse {

test_runner/batch_others/test_ancestor_branch.py

@@ -6,7 +6,6 @@ from fixtures.neon_fixtures import NeonEnv, NeonEnvBuilder, NeonPageserverApiExc
 #
 # Create ancestor branches off the main branch.
 #
-# @pytest.mark.parametrize("x", range(20))
 def test_ancestor_branch(neon_env_builder: NeonEnvBuilder):
     env = neon_env_builder.init_start()
 
@@ -22,7 +21,7 @@ def test_ancestor_branch(neon_env_builder: NeonEnvBuilder):
             'compaction_target_size': '4194304',
         })
 
-    # env.pageserver.safe_psql("failpoints flush-frozen-before-sync=sleep(10000)")
+    env.pageserver.safe_psql("failpoints flush-frozen-before-sync=sleep(10000)")
 
     pg_branch0 = env.postgres.create_start('main', tenant_id=tenant)
     branch0_cur = pg_branch0.connect().cursor()

test_runner/batch_others/test_branch_and_gc.py

@@ -1,4 +1,3 @@
-import pytest
 from fixtures.log_helper import log
 from fixtures.neon_fixtures import NeonEnv
 from fixtures.utils import lsn_from_hex
@@ -41,7 +40,6 @@ from fixtures.utils import lsn_from_hex
 # Because the delta layer D covering lsn1 is corrupted, creating a branch
 # starting from lsn1 should return an error as follows:
 #     could not find data for key ... at LSN ..., for request at LSN ...
-# @pytest.mark.skip("")
 def test_branch_and_gc(neon_simple_env: NeonEnv):
     env = neon_simple_env
 
@@ -58,9 +56,9 @@ def test_branch_and_gc(neon_simple_env: NeonEnv):
             'compaction_target_size': f'{1024 ** 3}',
 
             # tweak the default settings to allow quickly create image layers and L1 layers
-            'compaction_period': '10 s',
+            'compaction_period': '1 s',
             'compaction_threshold': '2',
-            'image_creation_threshold': '2',
+            'image_creation_threshold': '1',
 
             # set PITR interval to be small, so we can do GC
             'pitr_interval': '1 s'
@@ -79,9 +77,6 @@ def test_branch_and_gc(neon_simple_env: NeonEnv):
     lsn1 = main_cur.fetchone()[0]
     log.info(f'LSN1: {lsn1}')
 
-    # trigger a manual compaction
-    env.pageserver.safe_psql(f'''compact {tenant.hex} {timeline_main.hex}''')
-
     main_cur.execute('INSERT INTO foo SELECT FROM generate_series(1, 100000)')
     main_cur.execute('SELECT pg_current_wal_insert_lsn()')
     lsn2 = main_cur.fetchone()[0]

test_runner/batch_others/test_branching.py

@@ -44,7 +44,7 @@ def test_branching_with_pgbench(neon_simple_env: NeonEnv,
         log.info(f"Start a pgbench workload on pg {connstr}")
 
         pg_bin.run_capture(['pgbench', '-i', f'-s{scale}', connstr])
-        pg_bin.run_capture(['pgbench', '-T15', connstr])
+        pg_bin.run_capture(['pgbench', '-c10', '-T15', connstr])
 
     env.neon_cli.create_branch('b0', tenant_id=tenant)
     pgs: List[Postgres] = []
@@ -54,24 +54,12 @@ def test_branching_with_pgbench(neon_simple_env: NeonEnv,
     threads.append(threading.Thread(target=run_pgbench, args=(pgs[0], ), daemon=True))
     threads[-1].start()
 
-    thread_limit = 4
-
     for i in range(n_branches):
         # random a delay between [0, 5]
         delay = random.random() * 5
         time.sleep(delay)
         log.info(f"Sleep {delay}s")
 
-        # If the number of concurrent threads exceeds a threshold,
-        # wait for all the threads to finish before spawning a new one.
-        #
-        # Because tests defined in `batch_others` are run concurrently in CI,
-        # we want to avoid the situation that one test exhausts resources for other tests.
-        if len(threads) >= thread_limit:
-            for thread in threads:
-                thread.join()
-            threads = []
-
         if ty == "cascade":
             env.neon_cli.create_branch('b{}'.format(i + 1), 'b{}'.format(i), tenant_id=tenant)
         else: