DESCRIPTION: Parallelizes shard rebalancing and removes the bottlenecks
that previously blocked concurrent logical-replication moves.
These improvements reduce rebalance windows—particularly for clusters
with large reference tables and enable multiple shard transfers to run in parallel.
Motivation:
Citus’ shard rebalancer has some key performance bottlenecks:
**Sequential Movement of Reference Tables:**
Reference tables are often assumed to be small, but in real-world
deployments, they can grow significantly large. Previously, reference
table shards were transferred as a single unit, making the process
monolithic and time-consuming.
**No Parallelism Within a Colocation Group:**
Although Citus distributes data using colocated shards, shard
movements within the same colocation group were serialized. In
environments with hundreds of distributed tables colocated
together, this serialization significantly slowed down rebalance
operations.
**Excessive Locking:**
Rebalancer used restrictive locks and redundant logical replication
guards, further limiting concurrency.
The goal of this commit is to eliminate these inefficiencies and enable
maximum parallelism during rebalance, without compromising correctness
or compatibility. Parallelize shard rebalancing to reduce rebalance
time.
Feature Summary:
**1. Parallel Reference Table Rebalancing**
Each reference-table shard is now copied in its own background task.
Foreign key and other constraints are deferred until all shards are
copied.
For single shard movement without considering colocation a new
internal-only UDF '`citus_internal_copy_single_shard_placement`' is
introduced to allow single-shard copy/move operations.
Since this function is internal, we do not allow users to call it
directly.
**Temporary Hack to Set Background Task Context** Background tasks
cannot currently set custom GUCs like application_name before executing
internal-only functions. 'citus_rebalancer ...' statement as a prefix in
the task command. This is a temporary hack to label internal tasks until
proper GUC injection support is added to the background task executor.
**2. Changes in Locking Strategy**
- Drop the leftover replication lock that previously serialized shard
moves performed via logical replication. This lock was only needed when
we used to drop and recreate the subscriptions/publications before each
move. Since Citus now removes those objects later as part of the “unused
distributed objects” cleanup, shard moves via logical replication can
safely run in parallel without additional locking.
- Introduced a per-shard advisory lock to prevent concurrent operations
on the same shard while allowing maximum parallelism elsewhere.
- Change the lock mode in AcquirePlacementColocationLock from
ExclusiveLock to RowExclusiveLock to allow concurrent updates within the
same colocation group, while still preventing concurrent DDL operations.
**3. citus_rebalance_start() enhancements**
The citus_rebalance_start() function now accepts two new optional
parameters:
```
- parallel_transfer_colocated_shards BOOLEAN DEFAULT false,
- parallel_transfer_reference_tables BOOLEAN DEFAULT false
```
This ensures backward compatibility by preserving the existing behavior
and avoiding any disruption to user expectations and when both are set
to true, the rebalancer operates with full parallelism.
**Previous Rebalancer Behavior:**
`SELECT citus_rebalance_start(shard_transfer_mode := 'force_logical');`
This would:
Start a single background task for replicating all reference tables
Then, move all shards serially, one at a time.
```
Task 1: replicate_reference_tables()
↓
Task 2: move_shard_1()
↓
Task 3: move_shard_2()
↓
Task 4: move_shard_3()
```
Slow and sequential. Reference table copy is a bottleneck. Colocated
shards must wait for each other.
**New Parallel Rebalancer:**
```
SELECT citus_rebalance_start(
shard_transfer_mode := 'force_logical',
parallel_transfer_colocated_shards := true,
parallel_transfer_reference_tables := true
);
```
This would:
- Schedule independent background tasks for each reference-table shard.
- Move colocated shards in parallel, while still maintaining dependency
order.
- Defer constraint application until all reference shards are in place.
-
```
Task 1: copy_ref_shard_1()
Task 2: copy_ref_shard_2()
Task 3: copy_ref_shard_3()
→ Task 4: apply_constraints()
↓
Task 5: copy_shard_1()
Task 6: copy_shard_2()
Task 7: copy_shard_3()
↓
Task 8-10: move_shard_1..3()
```
Each operation is scheduled independently and can run as soon as
dependencies are satisfied.
This PR provides successful build against PG18Beta1. RuleUtils PR was
reviewed separately: #8010
## PG 18Beta1–related changes for building Citus
### TupleDesc / Attr layout
**What changed in PG:** Postgres consolidated the
`TupleDescData.attrs[]` array into a more compact representation. Direct
field access (tupdesc->attrs[i]) was replaced by the new
`TupleDescAttr()` API.
**Citus adaptation:** Everywhere we previously used
`tupdesc->attrs[...]`, we now call `TupleDescAttr(tupdesc, idx)` (or our
own `Attr()` macro) under a compatibility guard.
*
5983a4cffc
General Logic:
* Use `Attr(...)` in places where `columnar_version_compat.h` is
included. This avoids the need to sprinkle `#if PG_VERSION_NUM` guards
around each attribute access.
* Use `TupleDescAttr(tupdesc, i)` when the relevant PostgreSQL header is
already included and the additional macro indirection is unnecessary.
### Collation‐aware `LIKE`
**What changed in PG:** The `textlike` operator now requires an explicit
collation, to avoid ambiguous‐collation errors. Core code switched from
`DirectFunctionCall2(textlike, ...)` to
`DirectFunctionCall2Coll(textlike, DEFAULT_COLLATION_OID, ...)`.
**Citus adaptation:** In `remote_commands.c` and any other LIKE call, we
now use `DirectFunctionCall2Coll(textlike, DEFAULT_COLLATION_OID, ...)`
and `#include <utils/pg_collation.h>`.
*
85b7efa1cd
### Columnar storage API
* Adapt `columnar_relation_set_new_filelocator` (and related init
routines) for PG 18’s revised SMGR and storage-initialization hooks.
* Pull in the new headers (`explain_format.h`,
`columnar_version_compat.h`) so the columnar module compiles cleanly
against PG 18.
- heap_modify_tuple + heap_inplace_update only exist on PG < 18; on PG18
the in-place helper was removed upstream
-
a07e03fd8f
### OpenSSL / TLS integration
**What changed in PG:** Moved from the legacy `SSL_library_init()` to
`OPENSSL_init_ssl(OPENSSL_INIT_LOAD_CONFIG, NULL)`, updated certificate
API calls (`X509_getm_notBefore`, `X509_getm_notAfter`), and
standardized on `TLS_method()`.
**Citus adaptation:** We now `#include <openssl/opensslv.h>` and use
`#if OPENSSL_VERSION_NUMBER >= 0x10100000L` to choose between`
OPENSSL_init_ssl()` or `SSL_library_init()`, and wrap`
X509_gmtime_adj()` calls around the new accessor functions.
*
6c66b7443c
### Adapt `ExtractColumns()` to the new PG-18 `expandRTE()` signature
PostgreSQL 18
80feb727c8
added a fourth argument of type `VarReturningType` to `expandRTE()`, so
calls that used the old 7-parameter form no longer compile. This patch:
* Wraps the `expandRTE(...)` call in a `#if PG_VERSION_NUM >= 180000`
guard.
* On PG 18+ passes the new `VAR_RETURNING_DEFAULT` argument before
`location`.
* On PG 15–17 continues to call the original 7-arg form.
* Adds the necessary includes (`parser/parse_relation.h` for `expandRTE`
and `VarReturningType`, and `pg_version_constants.h` for
`PG_VERSION_NUM`).
### Adapt `ExecutorStart`/`ExecutorRun` hooks to PG-18’s new signatures
PostgreSQL 18
525392d572
changed the signatures of the executor hooks:
* `ExecutorStart_hook` now returns `bool` instead of `void`, and
* `ExecutorRun_hook` drops its old `run_once` argument.
This patch preserves Citus’s existing hook logic by:
1. **Adding two adapter functions** under `#if PG_VERSION_NUM >=
PG_VERSION_18`:
* `citus_executor_start_adapter(QueryDesc *queryDesc, int eflags)`
Calls the old `CitusExecutorStart(queryDesc, eflags)` and then returns
`true` to satisfy the new hook’s `bool` return type.
* `citus_executor_run_adapter(QueryDesc *queryDesc, ScanDirection
direction, uint64 count)`
Calls the old `CitusExecutorRun(queryDesc, direction, count, true)`
(passing `true` for the dropped `run_once` argument), and returns
`void`.
2. **Installing the adapters** in `_PG_init()` instead of the original
hooks when building against PG 18+:
```c
#if PG_VERSION_NUM >= PG_VERSION_18
ExecutorStart_hook = citus_executor_start_adapter;
ExecutorRun_hook = citus_executor_run_adapter;
#else
ExecutorStart_hook = CitusExecutorStart;
ExecutorRun_hook = CitusExecutorRun;
#endif
```
### Adapt to PG-18’s removal of the “run\_once” flag from
ExecutorRun/PortalRun
PostgreSQL commit
[[3eea7a0](3eea7a0c97)
rationalized the executor’s parallelism logic by moving the “execute a
plan only once” check into `ExecutePlan()` itself and dropping the old
`bool run_once` argument from the public APIs:
```diff
- void ExecutorRun(QueryDesc *queryDesc,
- ScanDirection direction,
- uint64 count,
- bool run_once);
+ void ExecutorRun(QueryDesc *queryDesc,
+ ScanDirection direction,
+ uint64 count);
```
(and similarly for `PortalRun()`).
To stay compatible across PG 15–18, Citus now:
1. **Updates all internal calls** to `ExecutorRun(...)` and
`PortalRun(...)`:
* On PG 18+, use the new three-argument form (`ExecutorRun(qd, dir,
count)`).
* On PG 15–17, keep the old four-arg form (`ExecutorRun(qd, dir, count,
true)`) under a `#if PG_VERSION_NUM < 180000` guard.
2. **Guards the dispatcher hooks** via the adapter functions (from the
earlier patch) so that Citus’s executor hooks continue to work under
both the old and new signatures.
### Adapt to PG-18’s shortened PortalRun signature
PostgreSQL 18’s refactoring (see commit
[3eea7a0](3eea7a0c97))
also removed the old run_once and alternate‐dest arguments from the
public PortalRun() API. The signature changed from:
```diff
- bool PortalRun(Portal portal,
- long count,
- bool isTopLevel,
- bool run_once,
- DestReceiver *dest,
- DestReceiver *altdest,
- QueryCompletion *qc);
+ bool PortalRun(Portal portal,
+ long count,
+ bool isTopLevel,
+ DestReceiver *dest,
+ DestReceiver *altdest,
+ QueryCompletion *qc);
```
To support both versions in Citus, we:
1. **Version-guard each call** to `PortalRun()`:
* **On PG 18+** invoke the new 6-argument form.
* **On PG 15–17** fall back to the legacy 7-argument form, passing
`true` for `run_once`.
### Add support for PG-18’s new `plansource` argument in
`PortalDefineQuery`**
PostgreSQL 18 extended the `PortalDefineQuery` API to carry a
`CachedPlanSource *plansource` pointer so that the portal machinery can
track cached‐plan invalidation (as introduced alongside deferred-locking
in commit
525392d572.
To remain compatible across PG 15–18, Citus now wraps its calls under a
version guard:
```diff
- PortalDefineQuery(portal, NULL, sql, commandTag, plantree_list, NULL);
+#if PG_VERSION_NUM >= 180000
+ /* PG 18+: seven-arg signature (adds plansource) */
+ PortalDefineQuery(
+ portal,
+ NULL, /* no prepared-stmt name */
+ sql, /* the query text */
+ commandTag, /* the CommandTag */
+ plantree_list, /* List of PlannedStmt* */
+ NULL, /* no CachedPlan */
+ NULL /* no CachedPlanSource */
+ );
+#else
+ /* PG 15–17: six-arg signature */
+ PortalDefineQuery(
+ portal,
+ NULL, /* no prepared-stmt name */
+ sql, /* the query text */
+ commandTag, /* the CommandTag */
+ plantree_list, /* List of PlannedStmt* */
+ NULL /* no CachedPlan */
+ );
+#endif
```
### Adapt ExecInitRangeTable() calls to PG-18’s new signature
PostgreSQL commit
[cbc127917e04a978a788b8bc9d35a70244396d5b](cbc127917e)
overhauled the planner API for range‐table initialization:
**PG 18+**: added a fourth `Bitmapset *unpruned_relids` argument to
support deferred partition pruning
In Citus’s `create_estate_for_relation()` (in `columnar_metadata.c`), we
now wrap the call in a compile‐time guard so that the code compiles
correctly on all supported PostgreSQL versions:
```
/* Prepare permission info on PG 16+ */
#if PG_VERSION_NUM >= PG_VERSION_16
List *perminfos = NIL;
addRTEPermissionInfo(&perminfos, rte);
#else
List *perminfos = NIL; /* unused on PG 15 */
#endif
/* Initialize the range table, with the right signature for each PG version */
#if PG_VERSION_NUM >= PG_VERSION_18
/* PG 18+: four‐arg signature (adds unpruned_relids) */
ExecInitRangeTable(
estate,
list_make1(rte),
perminfos,
NULL /* unpruned_relids: not used by columnar */
);
#elif PG_VERSION_NUM >= PG_VERSION_16
/* PG 16–17: three‐arg signature (permInfos) */
ExecInitRangeTable(
estate,
list_make1(rte),
perminfos
);
#else
/* PG 15: two‐arg signature */
ExecInitRangeTable(
estate,
list_make1(rte)
);
#endif
estate->es_output_cid = GetCurrentCommandId(true);
```
### Adapt `pgstat_report_vacuum()` to PG-18’s new timestamp argument
PostgreSQL commit
[[30a6ed0ce4bb18212ec38cdb537ea4b43bc99b83](30a6ed0ce4)
extended the `pgstat_report_vacuum()` API by adding a `TimestampTz
start_time` parameter at the end so that the VACUUM statistics collector
can record when the operation began:
```diff
/* PG ≤17: four-arg signature */
- void pgstat_report_vacuum(Oid tableoid,
- bool shared,
- double num_live_tuples,
- double num_dead_tuples);
+/* PG ≥18: five-arg signature adds a start_time */
+ void pgstat_report_vacuum(Oid tableoid,
+ bool shared,
+ double num_live_tuples,
+ double num_dead_tuples,
+ TimestampTz start_time);
```
To support both versions, we now wrap the call in `columnar_tableam.c`
with a version guard, supplying `GetCurrentTimestamp()` for PG-18+:
```c
#if PG_VERSION_NUM >= 180000
/* PG 18+: include start_timestamp */
pgstat_report_vacuum(
RelationGetRelid(rel),
rel->rd_rel->relisshared,
Max(new_live_tuples, 0), /* live tuples */
0, /* dead tuples */
GetCurrentTimestamp() /* start time */
);
#else
/* PG 15–17: original signature */
pgstat_report_vacuum(
RelationGetRelid(rel),
rel->rd_rel->relisshared,
Max(new_live_tuples, 0), /* live tuples */
0 /* dead tuples */
);
#endif
```
### Adapt `ExecuteTaskPlan()` to PG-18’s expanded `CreateQueryDesc()`
signature
PostgreSQL 18 changed `CreateQueryDesc()` from an eight-argument to a
nine-argument call by inserting a `CachedPlan *cplan` parameter
immediately after the `PlannedStmt *plannedstmt` argument (see commit
525392d572).
To remain compatible with PG 15–17, Citus now wraps its invocation in
`local_executor.c` with a version guard:
```diff
- /* PG15–17: eight-arg CreateQueryDesc without cached plan */
- QueryDesc *queryDesc = CreateQueryDesc(
- taskPlan, /* PlannedStmt *plannedstmt */
- queryString, /* const char *sourceText */
- GetActiveSnapshot(),/* Snapshot snapshot */
- InvalidSnapshot, /* Snapshot crosscheck_snapshot */
- destReceiver, /* DestReceiver *dest */
- paramListInfo, /* ParamListInfo params */
- queryEnv, /* QueryEnvironment *queryEnv */
- 0 /* int instrument_options */
- );
+#if PG_VERSION_NUM >= 180000
+ /* PG18+: nine-arg CreateQueryDesc with a CachedPlan slot */
+ QueryDesc *queryDesc = CreateQueryDesc(
+ taskPlan, /* PlannedStmt *plannedstmt */
+ NULL, /* CachedPlan *cplan (none) */
+ queryString, /* const char *sourceText */
+ GetActiveSnapshot(),/* Snapshot snapshot */
+ InvalidSnapshot, /* Snapshot crosscheck_snapshot */
+ destReceiver, /* DestReceiver *dest */
+ paramListInfo, /* ParamListInfo params */
+ queryEnv, /* QueryEnvironment *queryEnv */
+ 0 /* int instrument_options */
+ );
+#else
+ /* PG15–17: eight-arg CreateQueryDesc without cached plan */
+ QueryDesc *queryDesc = CreateQueryDesc(
+ taskPlan, /* PlannedStmt *plannedstmt */
+ queryString, /* const char *sourceText */
+ GetActiveSnapshot(),/* Snapshot snapshot */
+ InvalidSnapshot, /* Snapshot crosscheck_snapshot */
+ destReceiver, /* DestReceiver *dest */
+ paramListInfo, /* ParamListInfo params */
+ queryEnv, /* QueryEnvironment *queryEnv */
+ 0 /* int instrument_options */
+ );
+#endif
```
### Adapt `RelationGetPrimaryKeyIndex()` to PG-18’s new “deferrable\_ok”
flag
PostgreSQL commit
14e87ffa5c
added a new Boolean `deferrable_ok` parameter to
`RelationGetPrimaryKeyIndex()` so that the lock manager can defer
unique‐constraint locks when requested. The API changed from:
```c
RelationGetPrimaryKeyIndex(Relation relation)
```
to:
```c
RelationGetPrimaryKeyIndex(Relation relation, bool deferrable_ok)
```
```diff
diff --git a/src/backend/distributed/metadata/node_metadata.c
b/src/backend/distributed/metadata/node_metadata.c
index e3a1b2c..f4d5e6f 100644
--- a/src/backend/distributed/metadata/node_metadata.c
+++ b/src/backend/distributed/metadata/node_metadata.c
@@ -2965,8 +2965,18 @@
*/
- Relation replicaIndex =
index_open(RelationGetPrimaryKeyIndex(pgDistNode),
- AccessShareLock);
+ #if PG_VERSION_NUM >= PG_VERSION_18
+ /* PG 18+ adds a bool "deferrable_ok" parameter */
+ Relation replicaIndex =
+ index_open(
+ RelationGetPrimaryKeyIndex(pgDistNode, false),
+ AccessShareLock);
+ #else
+ Relation replicaIndex =
+ index_open(
+ RelationGetPrimaryKeyIndex(pgDistNode),
+ AccessShareLock);
+ #endif
ScanKeyInit(&scanKey[0], Anum_pg_dist_node_nodename,
BTEqualStrategyNumber, F_TEXTEQ, CStringGetTextDatum(nodeName));
```
```diff
diff --git a/src/backend/distributed/operations/node_protocol.c b/src/backend/distributed/operations/node_protocol.c
index e3a1b2c..f4d5e6f 100644
--- a/src/backend/distributed/operations/node_protocol.c
+++ b/src/backend/distributed/operations/node_protocol.c
@@ -746,7 +746,12 @@
if (!OidIsValid(idxoid))
{
- idxoid = RelationGetPrimaryKeyIndex(rel);
+ /* Determine the index OID of the primary key (PG18 adds a second parameter) */
+#if PG_VERSION_NUM >= PG_VERSION_18
+ idxoid = RelationGetPrimaryKeyIndex(rel, false);
+#else
+ idxoid = RelationGetPrimaryKeyIndex(rel);
+#endif
}
return idxoid;
```
Because Citus has always taken the lock immediately—just as the old
two-arg call did—we pass `false` to keep that same immediate-lock
behavior. Passing `true` would switch to deferred locking, which we
don’t want.
### Adapt `ExplainOnePlan()` to PG-18’s expanded API
PostgreSQL 18 extended
525392d572
the `ExplainOnePlan()` function to carry the `CachedPlan *` and
`CachedPlanSource *` pointers plus an explicit `query_index`, letting
the EXPLAIN machinery track plan‐source invalidation. The old signature:
```c
/* PG ≤17 */
void
ExplainOnePlan(PlannedStmt *plannedstmt,
IntoClause *into,
struct ExplainState *es,
const char *queryString,
ParamListInfo params,
QueryEnvironment *queryEnv,
const instr_time *planduration,
const BufferUsage *bufusage);
```
became, in PG 18:
```c
/* PG ≥18 */
void
ExplainOnePlan(PlannedStmt *plannedstmt,
CachedPlan *cplan,
CachedPlanSource *plansource,
int query_index,
IntoClause *into,
struct ExplainState *es,
const char *queryString,
ParamListInfo params,
QueryEnvironment *queryEnv,
const instr_time *planduration,
const BufferUsage *bufusage,
const MemoryContextCounters *mem_counters);
```
To compile under both versions, Citus now wraps each call in
`multi_explain.c` with:
```c
#if PG_VERSION_NUM >= PG_VERSION_18
/* PG 18+: pass NULL for the new cached‐plan fields and zero for query_index */
ExplainOnePlan(
plan, /* PlannedStmt *plannedstmt */
NULL, /* CachedPlan *cplan */
NULL, /* CachedPlanSource *plansource */
0, /* query_index */
into, /* IntoClause *into */
es, /* ExplainState *es */
queryString, /* const char *queryString */
params, /* ParamListInfo params */
NULL, /* QueryEnvironment *queryEnv */
&planduration,/* const instr_time *planduration */
(es->buffers ? &bufusage : NULL),
(es->memory ? &mem_counters : NULL)
);
#elif PG_VERSION_NUM >= PG_VERSION_17
/* PG 17: same as before, plus passing mem_counters if enabled */
ExplainOnePlan(
plan,
into,
es,
queryString,
params,
queryEnv,
&planduration,
(es->buffers ? &bufusage : NULL),
(es->memory ? &mem_counters : NULL)
);
#else
/* PG 15–16: original seven-arg form */
ExplainOnePlan(
plan,
into,
es,
queryString,
params,
queryEnv,
&planduration,
(es->buffers ? &bufusage : NULL)
);
#endif
```
### Adapt to the unified “index interpretation” API in PG 18 (commit
a8025f544854)
PostgreSQL commit
a8025f5448
generalized the old btree‐specific operator‐interpretation API into a
single “index interpretation” interface:
* **Renamed type**:
`OpBtreeInterpretation` → `OpIndexInterpretation`
* **Renamed function**:
`get_op_btree_interpretation(opno)` →
`get_op_index_interpretation(opno)`
* **Unified field**:
Each interpretation now carries `cmptype` instead of `strategy`.
To build cleanly on PG 18 while still supporting PG 15–17, Citus’s
shard‐pruning code now wraps these changes:
```c
#include "pg_version_constants.h"
#if PG_VERSION_NUM >= PG_VERSION_18
/* On PG 18+ the btree‐only APIs vanished; alias them to the new generic versions */
typedef OpIndexInterpretation OpBtreeInterpretation;
#define get_op_btree_interpretation(opno) get_op_index_interpretation(opno)
#define ROWCOMPARE_NE COMPARE_NE
#endif
/* … later, when checking an interpretation … */
OpBtreeInterpretation *interp =
(OpBtreeInterpretation *) lfirst(cell);
#if PG_VERSION_NUM >= PG_VERSION_18
/* use cmptype on PG 18+ */
if (interp->cmptype == ROWCOMPARE_NE)
#else
/* use strategy on PG 15–17 */
if (interp->strategy == ROWCOMPARE_NE)
#endif
{
/* … */
}
```
### Adapt `create_foreignscan_path()` for PG-18’s revised signature
PostgreSQL commit
e222534679
reordered and removed a couple of parameters in the FDW‐path builder:
* **PG 15–17 signature (11 args)**
```c
create_foreignscan_path(PlannerInfo *root,
RelOptInfo *rel,
PathTarget *target,
double rows,
Cost startup_cost,
Cost total_cost,
List *pathkeys,
Relids required_outer,
Path *fdw_outerpath,
List *fdw_restrictinfo,
List *fdw_private);
```
* **PG 18+ signature (9 args)**
```c
create_foreignscan_path(PlannerInfo *root,
RelOptInfo *rel,
PathTarget *target,
double rows,
int disabled_nodes,
Cost startup_cost,
Cost total_cost,
Relids required_outer,
Path *fdw_outerpath,
List *fdw_private);
```
To support both, Citus now defines a compatibility macro in
`pg_version_compat.h`:
```c
#include "nodes/bitmapset.h" /* for Relids */
#include "nodes/pg_list.h" /* for List */
#include "optimizer/pathnode.h" /* for create_foreignscan_path() */
#if PG_VERSION_NUM >= PG_VERSION_18
/* PG18+: drop pathkeys & fdw_restrictinfo, add disabled_nodes */
#define create_foreignscan_path_compat(a, b, c, d, e, f, g, h, i, j, k) \
create_foreignscan_path( \
(a), /* root */ \
(b), /* rel */ \
(c), /* target */ \
(d), /* rows */ \
(0), /* disabled_nodes (unused by Citus) */ \
(e), /* startup_cost */ \
(f), /* total_cost */ \
(g), /* required_outer */ \
(h), /* fdw_outerpath */ \
(k) /* fdw_private */ \
)
#else
/* PG15–17: original signature */
#define create_foreignscan_path_compat(a, b, c, d, e, f, g, h, i, j, k) \
create_foreignscan_path( \
(a), (b), (c), (d), \
(e), (f), \
(g), (h), (i), (j), (k) \
)
#endif
```
Now every call to `create_foreignscan_path_compat(...)`—even in tests
like `fake_fdw.c`—automatically picks the correct argument list for
PG 15 through PG 18.
### Drop the obsolete bitmap‐scan hooks on PG 18+
PostgreSQL commit
c3953226a0
cleaned up the `TableAmRoutine` API by removing the two bitmap‐scan
callback slots:
* `scan_bitmap_next_block`
* `scan_bitmap_next_tuple`
Since those hook‐slots no longer exist in PG 18, Citus now wraps their
NULL‐initialization in a `#if PG_VERSION_NUM < PG_VERSION_18` guard. On
PG 15–17 we still explicitly set them to `NULL` (to satisfy the old
struct layout), and on PG 18+ we omit them entirely:
```c
#if PG_VERSION_NUM < PG_VERSION_18
/* PG 15–17 only: these fields were removed upstream in PG 18 */
.scan_bitmap_next_block = NULL,
.scan_bitmap_next_tuple = NULL,
#endif
```
### Adapt `vac_update_relstats()` invocation to PG-18’s new
“all\_frozen” argument
PostgreSQL commit
99f8f3fbbc
extended the `vac_update_relstats()` API by inserting a
`num_all_frozen_pages` parameter between the existing
`num_all_visible_pages` and `hasindex` arguments:
```diff
- /* PG ≤17: */
- void
- vac_update_relstats(Relation relation,
- BlockNumber num_pages,
- double num_tuples,
- BlockNumber num_all_visible_pages,
- bool hasindex,
- TransactionId frozenxid,
- MultiXactId minmulti,
- bool *frozenxid_updated,
- bool *minmulti_updated,
- bool in_outer_xact);
+ /* PG ≥18: adds num_all_frozen_pages */
+ void
+ vac_update_relstats(Relation relation,
+ BlockNumber num_pages,
+ double num_tuples,
+ BlockNumber num_all_visible_pages,
+ BlockNumber num_all_frozen_pages,
+ bool hasindex,
+ TransactionId frozenxid,
+ MultiXactId minmulti,
+ bool *frozenxid_updated,
+ bool *minmulti_updated,
+ bool in_outer_xact);
```
To compile cleanly on both PG 15–17 and PG 18+, Citus wraps its call in
a version guard and supplies a zero placeholder for the new field:
```c
#if PG_VERSION_NUM >= 180000
/* PG 18+: supply explicit “all_frozen” count */
vac_update_relstats(
rel,
new_rel_pages,
new_live_tuples,
new_rel_allvisible, /* allvisible */
0, /* all_frozen */
nindexes > 0,
newRelFrozenXid,
newRelminMxid,
&frozenxid_updated,
&minmulti_updated,
false /* in_outer_xact */
);
#else
/* PG 15–17: original signature */
vac_update_relstats(
rel,
new_rel_pages,
new_live_tuples,
new_rel_allvisible,
nindexes > 0,
newRelFrozenXid,
newRelminMxid,
&frozenxid_updated,
&minmulti_updated,
false /* in_outer_xact */
);
#endif
```
**Why all_frozen = 0?**
Columnar storage never embeds transaction IDs in its pages, so it never
needs to track “all‐frozen” pages the way a heap does. Setting both
allvisible and allfrozen to zero simply tells Postgres “there are no
pages with the visibility or frozen‐status bits set,” matching our
existing behavior.
This change ensures Citus’s VACUUM‐statistic updates work unmodified
across all supported Postgres versions.
DESCRIPTION: Drops PG14 support
1. Remove "$version_num" != 'xx' from configure file
2. delete all PG_VERSION_NUM = PG_VERSION_XX references in the code
3. Look at pg_version_compat.h file, remove all _compat functions etc
defined specifically for PGXX differences
4. delete all PG_VERSION_NUM >= PG_VERSION_(XX+1), PG_VERSION_NUM <
PG_VERSION_(XX+1) ifs in the codebase
5. delete ruleutils_xx.c file
6. cleanup normalize.sed file from pg14 specific lines
7. delete all alternative output files for that particular PG version,
server_version_ge variable helps here
DESCRIPTION: Adds support for issuing `CREATE`/`DROP` DATABASE commands
from worker nodes
With this commit, we allow issuing CREATE / DROP DATABASE commands from
worker nodes too.
As in #7278, this is not allowed when the coordinator is not added to
metadata because we don't ever sync metadata changes to coordinator
when adding coordinator to the metadata via
`SELECT citus_set_coordinator_host('<hostname>')`, or equivalently, via
`SELECT citus_add_node(<coordinator_node_name>, <coordinator_node_port>, 0)`.
We serialize database management commands by acquiring a Citus specific
advisory lock on the first primary worker node if there are any workers in the
cluster. As opposed to what we've done in https://github.com/citusdata/citus/pull/7278
for role management commands, we try to avoid from running into distributed deadlocks
as much as possible. This is because, while distributed deadlocks that can happen around
role management commands can be detected by Citus, this is not the case for database
management commands because most of them cannot be run inside in a transaction block.
In that case, Citus cannot even detect the distributed deadlock because the command is not
part of a distributed transaction at all, then the command execution might not return the
control back to the user for an indefinite amount of time.
This change adds a script to programatically group all includes in a
specific order. The script was used as a one time invocation to group
and sort all includes throught our formatted code. The grouping is as
follows:
- System includes (eg. `#include<...>`)
- Postgres.h (eg. `#include "postgres.h"`)
- Toplevel imports from postgres, not contained in a directory (eg.
`#include "miscadmin.h"`)
- General postgres includes (eg . `#include "nodes/..."`)
- Toplevel citus includes, not contained in a directory (eg. `#include
"citus_verion.h"`)
- Columnar includes (eg. `#include "columnar/..."`)
- Distributed includes (eg. `#include "distributed/..."`)
Because it is quite hard to understand the difference between toplevel
citus includes and toplevel postgres includes it hardcodes the list of
toplevel citus includes. In the same manner it assumes anything not
prefixed with `columnar/` or `distributed/` as a postgres include.
The sorting/grouping is enforced by CI. Since we do so with our own
script there are not changes required in our uncrustify configuration.
DESCRIPTION: Add infrastructure to run long running management operations in background
This infrastructure introduces the primitives of jobs and tasks.
A task consists of a sql statement and an owner. Tasks belong to a
Job and can depend on other tasks from the same job.
When there are either runnable or running tasks we would like to
make sure a bacgrkound task queue monitor process is running. A Task
could be in running state while there is actually no monitor present
due to a database restart or failover. Once the monitor starts it
will reset any running task to its runnable state.
To make sure only one background task queue monitor is ever running
at once it will acquire an advisory lock that self conflicts.
Once a task is done it will find all tasks depending on this task.
After checking that the task doesn't have unmet dependencies it will
transition the task from blocked to runnable state for the task to
be picked up on a subsequent task start.
Currently only one task can be running at a time. This can be
improved upon in later releases without changes to the higher level
API.
The initial goal for this background tasks is to allow a rebalance
to run in the background. This will be implemented in a subsequent PR.
DESCRIPTION:
This PR adds support for 'Deferred Drop' and robust 'Shard Cleanup' for Splits.
Common Infrastructure
This PR introduces new common infrastructure so as any operation that wants robust cleanup of resources can register with the cleaner and have the resources cleaned appropriately based on a specified policy. 'Shard Split' is the first consumer using this new infrastructure.
Note : We only support adding 'shards' as resources to be cleaned-up right now but the framework will be extended to support other resources in future.
Deferred Drop for Split
Deferred Drop Support ensures that shards undergoing split are not dropped inline as part of operation but dropped later when no active read queries are running on shard. This helps with :
Avoids any potential deadlock scenarios that can cause long running Split operation to rollback.
Avoids Split operation blocking writes and then getting blocked (due to running queries on the shard) when trying to drop shards.
Deferred drop is the new default behavior going forward.
Shard Cleaner Extension
Shard Cleaner is a background task responsible for deferred drops in case of 'Move' operations.
The cleaner has been extended to ensure robust cleanup of shards (dummy shards and split children) in case of a failure based on the new infrastructure mentioned above. The cleaner also handles deferred drop for 'Splits'.
TESTING:
New test ''citus_split_shard_by_split_points_deferred_drop' to test deferred drop support.
New test 'failure_split_cleanup' to test shard cleanup with failures in different stages.
Update 'isolation_blocking_shard_split and isolation_non_blocking_shard_split' for deferred drop.
Added non-deferred drop version of existing tests : 'citus_split_shard_no_deferred_drop' and 'citus_non_blocking_splits_no_deferred_drop'
Added create_distributed_table_concurrently which is nonblocking variant of create_distributed_table.
It bases on the split API which takes advantage of logical replication to support nonblocking split operations.
Co-authored-by: Marco Slot <marco.slot@gmail.com>
Co-authored-by: aykutbozkurt <aykut.bozkurt1995@gmail.com>
It turns out that create_distributed_table
and citus_move/copy_shard_placement does not
work well concurrently.
To fix that, we need to acquire a lock, which
sounds like a good use of colocation lock.
However, the current usage of colocation lock is
limited to higher level UDFs like rebalance_table_shards
etc. Those usage of lock is still useful, but
we cannot acquire the same lock on citus_move_shard_placement
etc. because the coordinator connects to itself to acquire
the lock. Hence, the high level UDF blocks itself.
To fix that, we use one more colocation lock, with the placements
are the main objects to consider.
This PR makes all of the features open source that were previously only
available in Citus Enterprise.
Features that this adds:
1. Non blocking shard moves/shard rebalancer
(`citus.logical_replication_timeout`)
2. Propagation of CREATE/DROP/ALTER ROLE statements
3. Propagation of GRANT statements
4. Propagation of CLUSTER statements
5. Propagation of ALTER DATABASE ... OWNER TO ...
6. Optimization for COPY when loading JSON to avoid double parsing of
the JSON object (`citus.skip_jsonb_validation_in_copy`)
7. Support for row level security
8. Support for `pg_dist_authinfo`, which allows storing different
authentication options for different users, e.g. you can store
passwords or certificates here.
9. Support for `pg_dist_poolinfo`, which allows using connection poolers
in between coordinator and workers
10. Tracking distributed query execution times using
citus_stat_statements (`citus.stat_statements_max`,
`citus.stat_statements_purge_interval`,
`citus.stat_statements_track`). This is disabled by default.
11. Blocking tenant_isolation
12. Support for `sslkey` and `sslcert` in `citus.node_conninfo`
Breaking down #5899 into smaller PR-s
This particular PR changes the way TRUNCATE acquires distributed locks on the relations it is truncating to use the LOCK command instead of lock_relation_if_exists. This has the benefit of using pg's recursive locking logic it implements for the LOCK command instead of us having to resolve relation dependencies and lock them explicitly. While this does not directly affect truncate, it will allow us to generalize this locking logic to then log different relations where the pg recursive locking will become useful (e.g. locking views).
This implementation is a bit more complex that it needs to be due to pg not supporting locking foreign tables. We can however, still lock foreign tables with lock_relation_if_exists. So for a command:
TRUNCATE dist_table_1, dist_table_2, foreign_table_1, foreign_table_2, dist_table_3;
We generate and send the following command to all the workers in metadata:
```sql
SEL citus.enable_ddl_propagation TO FALSE;
LOCK dist_table_1, dist_table_2 IN ACCESS EXCLUSIVE MODE;
SELECT lock_relation_if_exists('foreign_table_1', 'ACCESS EXCLUSIVE');
SELECT lock_relation_if_exists('foreign_table_2', 'ACCESS EXCLUSIVE');
LOCK dist_table_3 IN ACCESS EXCLUSIVE MODE;
SEL citus.enable_ddl_propagation TO TRUE;
```
Note that we need to alternate between the lock command and lock_table_if_exists in order to preserve the TRUNCATE order of relations.
When pg supports locking foreign tables, we will be able to massive simplify this logic and send a single LOCK command.
- [x] Add some more regression test coverage
- [x] Make sure returning works fine in case of
local execution + remote execution
(task->partiallyLocalOrRemote works as expected, already added tests)
- [x] Implement locking properly (and add isolation tests)
- [x] We do #shardcount round-trips on `SerializeNonCommutativeWrites`.
We made it a single round-trip.
- [x] Acquire locks for subselects on the workers & add isolation tests
- [x] Add a GUC to prevent modification from the workers, hence increase the
coordinator-only throughput
- The performance slightly drops (~%15), unless
`citus.allow_modifications_from_workers_to_replicated_tables`
is set to false
Before this commit, we always synced the metadata with superuser.
However, that creates various edge cases such as visibility errors
or self distributed deadlocks or complicates user access checks.
Instead, with this commit, we use the current user to sync the metadata.
Note that, `start_metadata_sync_to_node` still requires super user
because accessing certain metadata (like pg_dist_node) always require
superuser (e.g., the current user should be a superuser).
However, metadata syncing operations regarding the distributed
tables can now be done with regular users, as long as the user
is the owner of the table. A table owner can still insert non-sense
metadata, however it'd only affect its own table. So, we cannot do
anything about that.
* Columnar: introduce columnar storage API.
This new API is responsible for the low-level storage details of
columnar; translating large reads and writes into individual block
reads and writes that respect the page headers and emit WAL. It's also
responsible for the columnar metapage, resource reservations (stripe
IDs, row numbers, and data), and truncation.
This new API is not used yet, but will be used in subsequent
forthcoming commits.
* Columnar: add columnar_storage_info() for debugging purposes.
* Columnar: expose ColumnarMetadataNewStorageId().
* Columnar: always initialize metapage at creation time.
This avoids the complexity of dealing with tables where the metapage
has not yet been initialized.
* Columnar: columnar storage upgrade/downgrade UDFs.
Necessary upgrade/downgrade step so that new code doesn't see an old
metapage.
* Columnar: improve metadata.c comment.
* Columnar: make ColumnarMetapage internal to the storage API.
Callers should not have or need direct access to the metapage.
* Columnar: perform resource reservation using storage API.
* Columnar: implement truncate using storage API.
* Columnar: implement read/write paths with storage API.
* Columnar: add storage tests.
* Revert "Columnar: don't include stripe reservation locks in lock graph."
This reverts commit c3dcd6b9f8.
No longer needed because the columnar storage API takes care of
concurrency for resource reservation.
* Columnar: remove unnecessary lock when reserving.
No longer necessary because the columnar storage API takes care of
concurrent resource reservation.
* Add simple upgrade tests for storage/ branch
* fix multi_extension.out
Co-authored-by: Onur Tirtir <onurcantirtir@gmail.com>
* Fix problews with concurrent calls of DropMarkedShards
When trying to enable `citus.defer_drop_after_shard_move` by default it
turned out that DropMarkedShards was not safe to call concurrently.
This could especially cause big problems when also moving shards at the
same time. During tests it was possible to trigger a state where a shard
that was moved would not be available on any of the nodes anymore after
the move.
Currently DropMarkedShards is only called in production by the
maintenaince deamon. Since this is only a single process triggering such
a race is currently impossible in production settings. In future changes
we will want to call DropMarkedShards from other places too though.
* Add some isolation tests
Co-authored-by: Jelte Fennema <github-tech@jeltef.nl>
* Not take ShareUpdateExlusiveLock on pg_dist_transaction
We were taking ShareUpdateExlusiveLock on pg_dist_transaction during
recovery to prevent multiple recoveries happening concurrenly. VACUUM(
not FULL) also takes ShareUpdateExclusiveLock, and they can conflict. It
seems that VACUUM will skip the table if there is a conflicting lock
already taken unless it is doing the vacuum to prevent id wraparound, in
which case there can be a deadlock. I guess the deadlock happens if:
- VACUUM takes a lock on pg_dist_transaction and is done for id
wraparound problem
- The transaction in the maintenance tries to take a lock but
cannot as that conflicts with the lock acquired by VACUUM
- The transaction in the maintenance daemon has a very old xid hence
VACUUM cannot proceed.
If we take a row exclusive lock in transaction recovery then it wouldn't
conflict with VACUUM hence it could proceed so the deadlock would be
resolved. To prevent concurrent transaction recoveries happening, an
advisory lock is taken with ShareUpdateExlusiveLock as before.
* Use CITUS_OPERATIONS tag
* Remove unused executor codes
All of the codes of real-time executor. Some functions
in router executor still remains there because there
are common functions. We'll move them to accurate places
in the follow-up commits.
* Move GUCs to transaction mngnt and remove unused struct
* Update test output
* Get rid of references of real-time executor from code
* Warn if real-time executor is picked
* Remove lots of unused connection codes
* Removed unused code for connection restrictions
Real-time and router executors cannot handle re-using of the existing
connections within a transaction block.
Adaptive executor and COPY can re-use the connections. So, there is no
reason to keep the code around for applying the restrictions in the
placement connection logic.
We acquire distributed lock on all mx nodes for truncated
tables before actually doing truncate operation.
This is needed for distributed serialization of the truncate
command without causing a deadlock.
Fixes#10
This change creates a new UDF: master_modify_multiple_shards
Parameters:
modify_query: A simple DELETE or UPDATE query as a string.
The UDF is similar to the existing master_apply_delete_command UDF.
Basically, given the modify query, it prunes the shard list, re-constructs
the query for each shard and sends the query to the placements.
Depending on the value of citus.multi_shard_commit_protocol, the commit
can be done in one-phase or two-phase manner.
Limitations:
* It cannot be called inside a transaction block
* It only be called with simple operator expressions (like Single Shard Modify)
Sample Usage:
```
SELECT master_modify_multiple_shards(
'DELETE FROM customer_delete_protocol WHERE c_custkey > 500 AND c_custkey < 500');
```
All citusdb references in
- extension, binary names
- file headers
- all configuration name prefixes
- error/warning messages
- some functions names
- regression tests
are changed to be citus.
naisila
Muhammad Usama
Mehmet YILMAZ
Naisila Puka
Onur Tirtir
Nils Dijk
Emel Şimşek
Nitish Upreti
Önder Kalacı
Marco Slot
Jelte Fennema
gledis69
Gledis Zeneli
Önder Kalacı
jeff-davis
SaitTalhaNisanci
Hadi Moshayedi
Marco Slot
Philip Dubé
Murat Tuncer
velioglu
mehmet furkan şahin
Dimitri Fontaine
Burak Yucesoy
Eren Basak
eren
Jason Petersen