Merge branch 'master' into query_cancellation_for_index_backfills

Author: egladysh

.arclint

@@ -12,10 +12,11 @@
      "([.]txt$)",
      "(^[.]gitmodules$)",
      "(Makefile$)",
-     "(src/yb/common/ql_type[.]cc$)",
-     "(src/yb/gutil/valgrind[.]h$)",
-     "(^src/odyssey/)",
      "(^src/inline-thirdparty/)",
+     "(^src/lint/cpplint[.]py$)",
+     "(^src/odyssey/)",
+     "(^src/yb/common/ql_type[.]cc$)",
+     "(^src/yb/gutil/valgrind[.]h$)",
      "(^src/yb/integration-tests/upgrade-tests/builds.xml$)",
      "(^managed/src/main/resources/alert/alert_templates[.]yml$)",
      "(^managed/src/main/resources/alert/.*[.]template$)",
@@ -142,6 +143,7 @@
         "([.]hpp$)",
         "([.]h$)"
       ],
+      "bin": "src/lint/cpplint.py",
       "cpplint.header_guard_root_dir": "src",
       "cpplint.max_line_length": 100,
       "severity.rules": {

.github/workflows/yb-ctl-test.yml

@@ -36,6 +36,9 @@ jobs:
       - uses: actions/checkout@v2
 
       - name: Run yb-ctl-test
+        env:
+          # yb-ctl-test.sh copies only scripts/installation; python unit tests live under python/.
+          YB_CTL_REPO_ROOT: ${{ github.workspace }}
         run: |
           set -x
           tmpdir=$(mktemp -d)

buf.lock

@@ -0,0 +1,6 @@
+# Generated by buf. DO NOT EDIT.
+version: v2
+deps:
+  - name: buf.build/googleapis/googleapis
+    commit: c17df5b2beca46928cc87d5656bd5343
+    digest: b5:648a01e0170d4512dea7d564016165decd1ed6e34bef79fe54753e51ad7e27545709ad9157d7551270147d551155c595a2fb0bf5bb33b1c83040ddbce915c604

buf.yaml

@@ -0,0 +1,9 @@
+# Local workspace for Buf CLI / VS Code "Buf" extension (proto navigation, lint).
+# Yugabyte CMake uses -I src for imports like "yb/common/common.proto".
+version: v2
+modules:
+  - path: src
+  # Local imports like "common.proto" (no yb/ prefix).
+  - path: managed/node-agent/proto
+deps:
+  - buf.build/googleapis/googleapis

build-support/lint.py

@@ -32,8 +32,7 @@
 
 
 REPO_ROOT = Path(__file__).resolve().parent.parent
-DEFAULT_CPPLINT = (shutil.which("cpplint.py")
-                   or str(REPO_ROOT / "thirdparty/installed/bin/cpplint.py"))
+CPPLINT_PATH = REPO_ROOT / "src" / "lint" / "cpplint.py"
 
 
 @dataclass
@@ -93,7 +92,7 @@ def _changed_files(base: str | None = None) -> list[str]:
 
     If ``base`` is provided, it is used directly as the diff base. Otherwise ``@{upstream}`` is
     used. If the current branch has no ``@{upstream}`` configured, the caller must pass
-    ``--rev`` explicitly — we do not fall back to ``origin/master`` because on stable release
+    ``--rev`` explicitly -- we do not fall back to ``origin/master`` because on stable release
     branches that would include a huge unrelated fileset.
     """
     if base is not None:
@@ -203,7 +202,7 @@ def _cpplint_filter_arg(severity_rules: dict) -> str:
 
 def run_cpplint(linter: Linter, files: list[str]) -> list[Message]:
     """Run Google's cpplint.py on ``files``."""
-    binary = DEFAULT_CPPLINT
+    binary = CPPLINT_PATH
     if not os.access(binary, os.X_OK):
         print(f"[lint] skipping {linter.name}: cpplint.py not executable at {binary}",
               file=sys.stderr)

build-support/tsan-suppressions.txt

@@ -187,3 +187,15 @@ race:k5_once
 race:od_logger_format
 race:od_stat_query_end
 race:od_route_pool_stat
+
+# pg_documentdb_gw_host bundles Rust code (tokio runtime, Rust stdlib, the
+# documentdb_gateway crate, and our bgworker glue) that is not compiled with
+# -Zsanitizer=thread (requires nightly Rust). TSAN cannot track the Rust
+# atomics that tokio uses for cross-thread task handoff and reports false
+# positive data races. Suppress the Rust call paths specifically so races
+# in C++ code that interacts with the gateway are still caught.
+race:tokio::
+race:std::fs::
+race:std::sys::
+race:documentdb_gateway::
+race:pg_documentdb_gw_host::

docs/content/stable/reference/configuration/yb-tserver.md

@@ -1087,8 +1087,7 @@ The number of tablet replicas that each core on a YB-TServer can support.
 
 {{% tags/wrap %}}
 
-
-Default: 1024 * (7/10) (corresponding to an overhead of roughly 700 KiB per tablet)
+Default: `1462`
 {{% /tags/wrap %}}
 
 The number of tablet replicas that each GiB reserved by YB-TServers for tablet overheads can support.

docs/content/stable/releases/yba-releases/v2025.1.md

@@ -49,12 +49,14 @@ Before upgrading, review the information in [Prepare to upgrade YugabyteDB Anywh
 ### Improvements
 
 * Enables setting priority class in Helm chart deployments. PLAT-19874
+* Enables viewing details of in-use image bundles in a view-only mode. PLAT-18642
 * Adds support for Python versions 3.6 and above, including 3.12, by installing `setuptools` and `pywheel` using `.whl` files and introducing a new flag `use_system_python` for system python usage. Ends script if `venv` isn't settable. PLAT-16111
 * Enables runtime modification of `yb.support_bundle.application_logs_regex_pattern`. PLAT-19369
 * Enables to configure `application_logs_sdf_pattern` and `application_logs_regex_pattern` at runtime for improved flexibility in generating support bundles, and allows custom prefix setting for support bundle filenames. PLAT-19625
 * Enables custom labels on Helm metadata via the `commonLabels` field in Helm overrides for better resource management. PLAT-19491
 * Adds Azure IAM integration in the storage configuration UI. PLAT-19502
 * Enables Azure AD Workload identity for more secure backups and restores. PLAT-8708
+* Increases the default timeout for `SetUniverseReplicationEnabled` RPC to 2 minutes. PLAT-20031
 
 ### Bug fixes
 
@@ -77,6 +79,15 @@ Before upgrading, review the information in [Prepare to upgrade YugabyteDB Anywh
 * Resolves error when re-adding previously deleted AZs to a provider configuration. PLAT-19457,PLAT-19003
 * Adjusts a configuration data type name from `Regex` to `Raw String` to ensure correctness. PLAT-19865
 * Enables customization of `tserver` liveness probes in the Helm chart. PLAT-18285
+* Adds check to stop Azure universes from using clockbound to prevent creation failures. PLAT-18912
+* Upgrades Go dependencies to patch vulnerabilities across multiple components. PLAT-19147
+* Blocks editing of provider settings that alter multi-AZ status during migrations. PLAT-19240
+* Upgrades Postgres to version 14.22, fixing multiple CVEs. PLAT-19701
+* Corrects the max process limit for YBA-provisioned universes. PLAT-20058
+* Enables correct updating of SSH keys in Terraform when modifying related fields. PLAT-20092
+* Ensures the 'yugabyte_upgrade' user is utilized during PostgreSQL upgrades in TLS-enabled universes, preventing upgrade failures. PLAT-20145
+* Fixes MIME type mismatch for platform configuration, ensuring UI compatibility with strict security policies. PLAT-20147
+* Ensures reliability in checking systemd units for versions earlier than 245. PLAT-20243
 
 ## v2025.1.3.2 - February 26, 2026 {#v2025.1.3.2}
 
@@ -141,7 +152,6 @@ This is a YugabyteDB-only release, with no changes to YugabyteDB Anywhere.
 * Upgraded Prometheus client library to eliminate warnings caused by duplicate JVM metrics. PLAT-17918
 * Simplified systemd handling by removing sudo check and installing service during node-agent installation. PLAT-18881
 * Introduces a runtime configuration to allow local user login even when Single Sign-On (SSO) is enabled on YugabyteDB Admin (YBA). This feature is enabled by default, but users have the option to disable it. PLAT-19204
-* Increases the default timeout for `SetUniverseReplicationEnabled` RPC to 2 minutes. PLAT-20031
 
 ### Bug fixes
 
@@ -176,15 +186,6 @@ This is a YugabyteDB-only release, with no changes to YugabyteDB Anywhere.
 * Refines common label escape hatch to prevent modification of default labels, thereby enhancing stability during universe creation. PLAT-19099
 * Fixes node agent installation failure for CSPs on Ubuntu. PLAT-19023
 * Fixed Helm formatting issue that occurred when using AWS/GCP exporter in otel-template. PLAT-19169
-* Adds check to stop Azure universes from using clockbound to prevent creation failures. PLAT-18912
-* Upgrades Go dependencies to patch vulnerabilities across multiple components. PLAT-19147
-* Blocks editing of provider settings that alter multi-AZ status during migrations. PLAT-19240
-* Upgrades Postgres to version 14.22, fixing multiple CVEs. PLAT-19701
-* Corrects the max process limit for YBA-provisioned universes. PLAT-20058
-* Enables correct updating of SSH keys in Terraform when modifying related fields. PLAT-20092
-* Ensures the 'yugabyte_upgrade' user is utilized during PostgreSQL upgrades in TLS-enabled universes, preventing upgrade failures. PLAT-20145
-* Fixes MIME type mismatch for platform configuration, ensuring UI compatibility with strict security policies. PLAT-20147
-* Ensures reliability in checking systemd units for versions earlier than 245. PLAT-20243
 
 ## v2025.1.2.2 - December 17, 2025 {#v2025.1.2.2}
 

docs/content/stable/yugabyte-platform/back-up-restore-universes/disaster-recovery/_index.md

@@ -51,6 +51,9 @@ Video: [Disaster Recovery With xCluster DR and Two Cloud Regions](https://www.yo
     href="disaster-recovery-setup/"
     icon="fa-thin fa-umbrella">}}
 
+{{</index/block>}}
+{{<index/block>}}
+
   {{<index/item
     title="Unplanned failover"
     body="Fail over to the DR replica in case of an unplanned outage."
@@ -63,12 +66,6 @@ Video: [Disaster Recovery With xCluster DR and Two Cloud Regions](https://www.yo
     href="disaster-recovery-switchover/"
     icon="fa-thin fa-toggle-on">}}
 
-  {{<index/item
-    title="Add and remove tables and indexes"
-    body="Perform DDL changes to databases in replication."
-    href="disaster-recovery-tables/"
-    icon="fa-thin fa-plus-minus">}}
-
 {{</index/block>}}
 
 ## Schema change modes
@@ -81,6 +78,8 @@ xCluster DR can be set up to perform schema changes in the following ways:
 | [Semi-automatic](#semi-automatic-mode) | Compared to manual mode, provides operationally simpler setup and management of replication, and fewer steps for performing DDL changes. | v2025.1.0 | v2025.2.1 |
 | [Manual](#manual-mode) | Deprecated. Manual setup and management of replication. DDL changes require manually updating the xCluster configuration. | v2024.2 | v2025.1 |
 
+For information on how each mode behaves at the database layer, refer to [xCluster replication](../../../architecture/docdb-replication/async-replication/).
+
 ### Automatic mode
 
 In automatic mode, all table and index-level schema changes made to the DR primary universe are automatically replicated to the DR replica.
@@ -89,7 +88,7 @@ You don't need to make any changes to the DR configuration.
 
 Automatic mode is recommended for all new DR configurations. When possible, you should delete existing DR configurations and re-create them using automatic mode to reduce the operational burden of DDL changes.
 
-Automatic mode is used for any xCluster DR configuration when both DR primary and replica are running YugabyteDB {{<release "2025.2.1">}} or later. For earlier versions, semi-automatic mode is used.
+By default, automatic mode is used for new DR configurations when both DR primary and replica are running YugabyteDB v2025.2.1 or later. If automatic mode is not used when you create a DR configuration, check the [runtime parameters](#schema-change-mode-runtime-configuration) are set correctly on DR primary.
 
 ### Semi-automatic mode
 
@@ -100,7 +99,7 @@ In this mode, table and index-level schema changes must be performed in the same
 
 You don't need to make any changes to the DR configuration.
 
-Semi-automatic mode is used for any xCluster DR configuration when both DR primary and replica are running YugabyteDB {{<release "2024.1.3">}} to {{<release "2025.2.0">}}.
+By default, semi-automatic mode is used for new DR configurations when both DR primary and replica are running YugabyteDB {{<release "2024.1.3">}} to {{<release "2025.2.0">}}. If semi-automatic mode is not used when you create a DR configuration, check the [runtime parameters](#schema-change-mode-runtime-configuration) are set correctly on DR primary.
 
 {{<lead link="https://www.youtube.com/watch?v=vYyn2OUSZFE">}}
 To learn more, watch [Simplified schema management with xCluster DB Scoped](https://www.youtube.com/watch?v=vYyn2OUSZFE)
@@ -116,7 +115,30 @@ In manual mode, table and index-level schema changes must be performed on the DR
 
 The exact sequence of these operations for each type of schema change (DDL) is described in [Manage tables and indexes](./disaster-recovery-tables/).
 
-Manual mode is used for any xCluster DR configuration when both DR primary and replica are running YugabyteDB {{<release "2024.1.2">}} or earlier.
+By default, Manual mode is used for DR configurations when both DR primary and replica are running YugabyteDB {{<release "2024.1.2">}} or earlier.
+
+### Upgrading schema change mode
+
+If you are running manual mode and you are running a compatible version of YugabyteDB, it is recommended that you upgrade to semi-automatic or automatic mode. Likewise, if you are running semi-automatic mode, it is recommended that you upgrade to automatic mode.
+
+To upgrade schema change mode for an existing DR configuration:
+
+1. [Remove disaster recovery](./disaster-recovery-setup/#remove-disaster-recovery).
+1. [Set up a new DR configuration](./disaster-recovery-setup/#set-up-disaster-recovery).
+
+### Schema change mode runtime configuration
+
+When you create a new DR configuration, YugabyteDB Anywhere selects semi-automatic or automatic mode based on the YugabyteDB version _and_ the following universe runtime parameters:
+
+- **Enable xCluster DR Semi-automatic Mode** (config key `yb.xcluster.db_scoped.creationEnabled`)
+
+- **Enable xCluster DR Automatic Mode** (config key `yb.xcluster.db_scoped.automatic_ddl.creationEnabled`)
+
+For new DR configurations, YugabyteDB Anywhere uses automatic mode when both universes are running YugabyteDB {{<release "2025.2.1">}} or later and _both parameters_ are true (the default in {{<release "2025.2.1">}} and later) on the DR primary.
+
+If **Enable xCluster DR Automatic Mode** is `false` while **Enable xCluster DR Semi-automatic Mode** remains `true`, new configurations use semi-automatic mode for supported versions.
+
+To set these options and their default behavior, refer to [Manage runtime configuration settings](../../administer-yugabyte-platform/manage-runtime-config/).
 
 ## Upgrading universes in DR
 

docs/content/stable/yugabyte-platform/back-up-restore-universes/disaster-recovery/disaster-recovery-setup.md

@@ -24,7 +24,7 @@ Create two universes, the DR primary universe which will serve reads and writes,
 
 Ensure the universes have the following characteristics:
 
-- Both universes are running the same version of YugabyteDB (v2.18.0.0 or later).
+- Both universes are running the same version of YugabyteDB (v2.18.0.0 or later). The schema change mode used by your new configuration depends on the version you are running, and runtime configuration. Refer to [Schema change modes](../#schema-change-modes).
 - Both universes have the same [encryption in transit](../../../security/enable-encryption-in-transit/) settings. Encryption in transit is recommended, and you should create the DR primary and DR replica universes with TLS enabled.
 - They can be backed up and restored using the same [storage configuration](../../configure-backup-storage/).
 - They have enough disk space to support storage of write-ahead logs (WALs) in case of a network partition or a temporary outage of the DR replica universe. During these cases, WALs will continue to write until replication is restored. Consider sizing your disk according to your ability to respond and recover from network or other infrastructure outages.