Change timeline_offloading setting default to true (#10760)

This changes the default value of the `timeline_offloading` pageserver
and tenant configs to true, now that offloading has been rolled out
without problems.

There is also a small fix in the tenant config merge function, where we
applied the `lazy_slru_download` value instead of `timeline_offloading`.

Related issue: https://github.com/neondatabase/cloud/issues/21353

libs/pageserver_api/src/config.rs

@@ -493,7 +493,7 @@ impl Default for ConfigToml {
                 NonZeroUsize::new(DEFAULT_MAX_VECTORED_READ_BYTES).unwrap(),
             )),
             image_compression: (DEFAULT_IMAGE_COMPRESSION),
-            timeline_offloading: false,
+            timeline_offloading: true,
             ephemeral_bytes_per_memory_kb: (DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB),
             l0_flush: None,
             virtual_file_io_mode: None,
@@ -624,7 +624,7 @@ impl Default for TenantConfigToml {
             image_creation_preempt_threshold: DEFAULT_IMAGE_CREATION_PREEMPT_THRESHOLD,
             lsn_lease_length: LsnLease::DEFAULT_LENGTH,
             lsn_lease_length_for_ts: LsnLease::DEFAULT_LENGTH_FOR_TS,
-            timeline_offloading: false,
+            timeline_offloading: true,
             wal_receiver_protocol_override: None,
             rel_size_v2_enabled: None,
             gc_compaction_enabled: DEFAULT_GC_COMPACTION_ENABLED,

pageserver/src/tenant/config.rs

@@ -466,7 +466,7 @@ impl TenantConfOpt {
                 .lsn_lease_length_for_ts
                 .unwrap_or(global_conf.lsn_lease_length_for_ts),
             timeline_offloading: self
-                .lazy_slru_download
+                .timeline_offloading
                 .unwrap_or(global_conf.timeline_offloading),
             wal_receiver_protocol_override: self
                 .wal_receiver_protocol_override

test_runner/regress/test_attach_tenant_config.py

@@ -175,7 +175,7 @@ def test_fully_custom_config(positive_env: NeonEnv):
         "image_layer_creation_check_threshold": 1,
         "lsn_lease_length": "1m",
         "lsn_lease_length_for_ts": "5s",
-        "timeline_offloading": True,
+        "timeline_offloading": False,
         "wal_receiver_protocol_override": {
             "type": "interpreted",
             "args": {"format": "bincode", "compression": {"zstd": {"level": 1}}},

test_runner/regress/test_timeline_archive.py

@@ -139,9 +139,9 @@ def test_timeline_archive(neon_env_builder: NeonEnvBuilder, shard_count: int):
 
 @pytest.mark.parametrize("manual_offload", [False, True])
 def test_timeline_offloading(neon_env_builder: NeonEnvBuilder, manual_offload: bool):
-    if not manual_offload:
-        # (automatic) timeline offloading defaults to false for now
-        neon_env_builder.pageserver_config_override = "timeline_offloading = true"
+    if manual_offload:
+        # (automatic) timeline offloading defaults to true
+        neon_env_builder.pageserver_config_override = "timeline_offloading = false"
 
     env = neon_env_builder.init_start()
     ps_http = env.pageserver.http_client()
@@ -396,8 +396,6 @@ def test_timeline_archival_chaos(neon_env_builder: NeonEnvBuilder):
     with tenant migrations and timeline deletions.
     """
 
-    # Offloading is off by default at time of writing: remove this line when it's on by default
-    neon_env_builder.pageserver_config_override = "timeline_offloading = true"
     neon_env_builder.storage_controller_config = {"heartbeat_interval": "100msec"}
     neon_env_builder.enable_pageserver_remote_storage(s3_storage())
 
@@ -994,8 +992,6 @@ def test_timeline_offload_race_unarchive(
     Ensure that unarchive and timeline offload don't race each other
     """
     # Regression test for issue https://github.com/neondatabase/neon/issues/10220
-    # (automatic) timeline offloading defaults to false for now
-    neon_env_builder.pageserver_config_override = "timeline_offloading = true"
 
     failpoint = "before-timeline-auto-offload"