Before this commit, shardPlacements were identified with shardId, nodeName
and nodeport. Instead of using nodeName and nodePort, we now use nodeId
since it apparently has performance benefits in several places in the
code.
Before this commit, Citus supported INSERT...SELECT queries with
ON CONFLICT or RETURNING clauses only for pushdownable ones, since
queries supported via coordinator were utilizing COPY infrastructure
of PG to send selected tuples to the target worker nodes.
After this PR, INSERT...SELECT queries with ON CONFLICT or RETURNING
clauses will be performed in two phases via coordinator. In the first
phase selected tuples will be saved to the intermediate table which
is colocated with target table of the INSERT...SELECT query. Note that,
a utility function to save results to the colocated intermediate result
also implemented as a part of this commit. In the second phase, INSERT..
SELECT query is directly run on the worker node using the intermediate
table as the source table.
The file handling the utility functions (DDL) for citus organically grew over time and became unreasonably large. This refactor takes that file and refactored the functionality into separate files per command. Initially modeled after the directory and file layout that can be found in postgres.
Although the size of the change is quite big there are barely any code changes. Only one two functions have been added for readability purposes:
- PostProcessIndexStmt which is extracted from PostProcessUtility
- PostProcessAlterTableStmt which is extracted from multi_ProcessUtility
A README.md has been added to `src/backend/distributed/commands` describing the contents of the module and every file in the module.
We need more documentation around the overloading of the COPY command, for now the boilerplate has been added for people with better knowledge to fill out.
With this commit, we all partitioned distributed tables with
replication factor > 1. However, we also have many restrictions.
In summary, we disallow all kinds of modifications (including DDLs)
on the partition tables. Instead, the user is allowed to run the
modifications over the parent table.
The necessity for such a restriction have two aspects:
- We need to acquire shard resource locks appropriately
- We need to handle marking partitions INVALID in case
of any failures. Note that, in theory, the parent table
should also become INVALID, which is too aggressive.
When a hash distributed table have a foreign key to a reference
table, there are few restrictions we have to apply in order to
prevent distributed deadlocks or reading wrong results.
The necessity to apply the restrictions arise from cascading
nature of foreign keys. When a foreign key on a reference table
cascades to a distributed table, a single operation over a single
connection can acquire locks on multiple shards of the distributed
table. Thus, any parallel operation on that distributed table, in the
same transaction should not open parallel connections to the shards.
Otherwise, we'd either end-up with a self-distributed deadlock or
read wrong results.
As briefly described above, the restrictions that we apply is done
by tracking the distributed/reference relation accesses inside
transaction blocks, and act accordingly when necessary.
The two main rules are as follows:
- Whenever a parallel distributed relation access conflicts
with a consecutive reference relation access, Citus errors
out
- Whenever a reference relation access is followed by a
conflicting parallel relation access, the execution mode
is switched to sequential mode.
There are also some other notes to mention:
- If the user does SET LOCAL citus.multi_shard_modify_mode
TO 'sequential';, all the queries should simply work with
using one connection per worker and sequentially executing
the commands. That's obviously a slower approach than Citus'
usual parallel execution. However, we've at least have a way
to run all commands successfully.
- If an unrelated parallel query executed on any distributed
table, we cannot switch to sequential mode. Because, the essense
of sequential mode is using one connection per worker. However,
in the presence of a parallel connection, the connection manager
picks those connections to execute the commands. That contradicts
with our purpose, thus we error out.
- COPY to a distributed table cannot be executed in sequential mode.
Thus, if we switch to sequential mode and COPY is executed, the
operation fails and there is currently no way of implementing that.
Note that, when the local table is not empty and create_distributed_table
is used, citus uses COPY internally. Thus, in those cases,
create_distributed_table() will also fail.
- There is a GUC called citus.enforce_foreign_key_restrictions
to disable all the checks. We added that GUC since the restrictions
we apply is sometimes a bit more restrictive than its necessary.
The user might want to relax those. Similarly, if you don't have
CASCADEing reference tables, you might consider disabling all the
checks.
- changes in ruleutils_11.c is reflected
- vacuum statement api change is handled. We now allow
multi-table vacuum commands.
- some other function header changes are reflected
- api conflicts between PG11 and earlier versions
are handled by adding shims in version_compat.h
- various regression tests are fixed due output and
functionality in PG1
- no change is made to support new features in PG11
they need to be handled by new commit
Postgres provides OS agnosting formatting macros for
formatting 64 bit numbers. Replaced %ld %lu with
INT64_FORMAT and UINT64_FORMAT respectively.
Also found some incorrect usages of formatting
flags and fixed them.
It's possible to build INSERT SELECT queries which include implicit
casts, currently we attempt to support these by adding explicit casts to
the SELECT query, but this sometimes crashes because we don't update all
nodes with the new types. (SortClauses, for instance)
This commit removes those explicit casts and passes an unmodified SELECT
query to the COPY executor (how we implement INSERT SELECT under the
scenes). In lieu of those cases, COPY has been given some extra logic to
inspect queries, notice that the types don't line up with the table it's
supposed to be inserting into, and "manually" casting every tuple before
sending them to workers.
When a NULL connection is provided to PQerrorMessage(), the
returned error message is a static text. Modifying that static
text, which doesn't necessarly be in a writeable memory, is
dangreous and might cause a segfault.
For partitioned tables, PostgreSQL opens partition and its partitions
in BeginCopyFrom and it expects its caller to close those relations.
However, we do not have quick access to opened relations and performing
special operations for partitioned tables isn't necessary in coordinator
node. Therefore before calling BeginCopyFrom, we change relkind of those
partitioned tables to RELKIND_RELATION. This prevents PostgreSQL to open
its partitions as well.
Adds support for PostgreSQL 10 by copying in the requisite ruleutils
and updating all API usages to conform with changes in PostgreSQL 10.
Most changes are fairly minor but they are numerous. One particular
obstacle was the change in \d behavior in PostgreSQL 10's psql; I had
to add SQL implementations (views, mostly) to mimic the pre-10 output.
Add a second implementation of INSERT INTO distributed_table SELECT ... that is used if
the query cannot be pushed down. The basic idea is to execute the SELECT query separately
and pass the results into the distributed table using a CopyDestReceiver, which is also
used for COPY and create_distributed_table. When planning the SELECT, we go through
planner hooks again, which means the SELECT can also be a distributed query.
EXPLAIN is supported, but EXPLAIN ANALYZE is not because preventing double execution was
a lot more complicated in this case.
So far citus used postgres' predicate proofing logic for shard
pruning, except for INSERT and COPY which were already optimized for
speed. That turns out to be too slow:
* Shard pruning for SELECTs is currently O(#shards), because
PruneShardList calls predicate_refuted_by() for every
shard. Obviously using an O(N) type algorithm for general pruning
isn't good.
* predicate_refuted_by() is quite expensive on its own right. That's
primarily because it's optimized for doing a single refutation
proof, rather than performing the same proof over and over.
* predicate_refuted_by() does not keep persistent state (see 2.) for
function calls, which means that a lot of syscache lookups will be
performed. That's particularly bad if the partitioning key is a
composite key, because without a persistent FunctionCallInfo
record_cmp() has to repeatedly look-up the type definition of the
composite key. That's quite expensive.
Thus replace this with custom-code that works in two phases:
1) Search restrictions for constraints that can be pruned upon
2) Use those restrictions to search for matching shards in the most
efficient manner available:
a) Binary search / Hash Lookup in case of hash partitioned tables
b) Binary search for equal clauses in case of range or append
tables without overlapping shards.
c) Binary search for inequality clauses, searching for both lower
and upper boundaries, again in case of range or append
tables without overlapping shards.
d) exhaustive search testing each ShardInterval
My measurements suggest that we are considerably, often orders of
magnitude, faster than the previous solution, even if we have to fall
back to exhaustive pruning.
All callers fetch a cache entry and extract/compute arguments for the
eventual FindShardInterval call, so it makes more sense to refactor
into that function itself; this solves the use-after-free bug, too.
- Break CheckShardPlacements into multiple functions (The most important
is MarkFailedShardPlacements), so that we can get rid of the global
CoordinatedTransactionUses2PC.
- Call MarkFailedShardPlacements in the router executor, so we mark
shards as invalid and stop using them while inside transaction blocks.
This enables proper transactional behaviour for copy and relaxes some
restrictions like combining COPY with single-row modifications. It
also provides the basis for relaxing restrictions further, and for
optionally allowing connection caching.
With this commit, we implemented some basic features of reference tables.
To start with, a reference table is
* a distributed table whithout a distribution column defined on it
* the distributed table is single sharded
* and the shard is replicated to all nodes
Reference tables follows the same code-path with a single sharded
tables. Thus, broadcast JOINs are applicable to reference tables.
But, since the table is replicated to all nodes, table fetching is
not required any more.
Reference tables support the uniqueness constraints for any column.
Reference tables can be used in INSERT INTO .. SELECT queries with
the following rules:
* If a reference table is in the SELECT part of the query, it is
safe join with another reference table and/or hash partitioned
tables.
* If a reference table is in the INSERT part of the query, all
other participating tables should be reference tables.
Reference tables follow the regular co-location structure. Since
all reference tables are single sharded and replicated to all nodes,
they are always co-located with each other.
Queries involving only reference tables always follows router planner
and executor.
Reference tables can have composite typed columns and there is no need
to create/define the necessary support functions.
All modification queries, master_* UDFs, EXPLAIN, DDLs, TRUNCATE,
sequences, transactions, COPY, schema support works on reference
tables as expected. Plus, all the pre-requisites associated with
distribution columns are dismissed.
So far placements were assigned an Oid, but that was just used to track
insertion order. It also did so incompletely, as it was not preserved
across changes of the shard state. The behaviour around oid wraparound
was also not entirely as intended.
The newly introduced, explicitly assigned, IDs are preserved across
shard-state changes.
The prime goal of this change is not to improve ordering of task
assignment policies, but to make it easier to reference shards. The
newly introduced UpdateShardPlacementState() makes use of that, and so
will the in-progress connection and transaction management changes.
Three changes here to get to true multi-statement, multi-relation DDL
transactions (same functionality pre-5.2, with benefits of atomicity):
1. Changed the multi-shard utility hook to always run (consistency
with router executor hook, removes ad-hoc "installed" boolean)
2. Change the global connection list in multi_shard_transaction to
instead be a hash; update related functions to operate on global
hash instead of local hash/global list
3. Remove check within DDL code to prevent subsequent DDL commands;
place unset/reset guard around call to ConnectToNode to permit
connecting to additional nodes after DDL transaction has begun
In addition, code has been added to raise an error if a ROLLBACK TO
SAVEPOINT is attempted (similar to router executor), and comprehensive
tests execute all multi-DDL scenarios (full success, user ROLLBACK, any
actual errors (say, duplicate index), partial failure (duplicate index
on one node but not others), partial COMMIT (one node fails), and 2PC
partial PREPARE (one node fails)). Interleavings with other commands
(DML, \copy) are similarly all covered.
Recent changes to DDL and transaction logic resulted in a "regression"
from the viewpoint of users. Previously, DDL commands were allowed in
multi-command transaction blocks, though they were not processed in any
actual transactional manner. We improved the atomicity of our DDL code,
but added a restriction that DDL commands themselves must not occur in
any BEGIN/END transaction block.
To give users back the original functionality (and improved atomicity)
we now keep track of whether a multi-command transaction has modified
data (DML) or schema (DDL). Interleaving the two modification types in
a single transaction is disallowed.
This first step simply permits a single DDL command in such a block,
admittedly an incomplete solution, but one which will permit us to add
full multi-DDL command support in a subsequent commit.
Fixes#513
This change modifies the DDL Propagation logic so that DDL queries
are propagated via 2-Phase Commit protocol. This way, failures during
the execution of distributed DDL commands will not leave the table in
an intermediate state and the pending prepared transactions can be
commited manually.
DDL commands are not allowed inside other transaction blocks or functions.
DDL commands are performed with 2PC regardless of the value of
`citus.multi_shard_commit_protocol` parameter.
The workflow of the successful case is this:
1. Open individual connections to all shard placements and send `BEGIN`
2. Send `SELECT worker_apply_shard_ddl_command(<shardId>, <DDL Command>)`
to all connections, one by one, in a serial manner.
3. Send `PREPARE TRANSCATION <transaction_id>` to all connections.
4. Sedn `COMMIT` to all connections.
Failure cases:
- If a worker problem occurs before sending of all DDL commands is finished, then
all changes are rolled back.
- If a worker problem occurs after all DDL commands are sent but not after
`PREPARE TRANSACTION` commands are finished, then all changes are rolled back.
However, if a worker node is failed, then the prepared transactions in that worker
should be rolled back manually.
- If a worker problem occurs during `COMMIT PREPARED` statements are being sent,
then the prepared transactions on the failed workers should be commited manually.
- If master fails before the first 'PREPARE TRANSACTION' is sent, then nothing is
changed on workers.
- If master fails during `PREPARE TRANSACTION` commands are being sent, then the
prepared transactions on workers should be rolled back manually.
- If master fails during `COMMIT PREPARED` or `ROLLBACK PREPARED` commands are being
sent, then the remaining prepared transactions on the workers should be handled manually.
This change also helps with #480, since failed DDL changes no longer mark
failed placements as inactive.
Fixes#463
OID of user-defined types may be different in master and worker nodes. This causes errors
while sending data between nodes with binary nodes. Because binary copy format adds OID
of the element if it is in an array. The code adding OID is in PostgreSQL code, therefore
we cannot change it. Instead we decided to use text format if we try to send array of
user-defined type.
Fixes#550, fixes#545
If table name contains special characters, it needs to be escaped. However in some cases,
we escape table name before appending shardId, which causes syntax error in the queries
sent to worker nodes. With this change we now append shardId before escaping table names.
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');
```
Now, we can copy to an append-partitioned distributed relation from
any worker node by providing master options such as;
COPY relation_name FROM file_path WITH (delimiter '|', master_host 'localhost', master_port 5432);
where master_port is optional and default is 5432.
This change renames the distributed transaction manager parameter from
citus.copy_transaction_manager to citus.multi_shard_commit_protocol.
Distributed transaction manager has been used only by the COPY on hash
partitioned tables but it can be used by upcoming features so, we needed
to rename so that its name do not contain a reference to COPY.
The change also includes renames like transaction_manager_options to
commit_protocol_options and TRANSACTION_MANAGER_1PC to COMMIT_PROTOCOL_1PC.
With this change, declaration of MultiShardCommitProtocol (was
CopyTransactionManager) is moved from multi_copy.c to multi_transaction.c.
So far we've always used libpq defaults when connecting to workers; bar
special environment variables being set that'll always be the user that
started the server. That's not desirable because it prevents using
users with fewer privileges.
Thus change the various APIs creating connections to workers to always
use usernames. That means:
1) MultiClientConnect() needs to, optionally, accept a username
2) GetOrEstablishConnection(), including the underlying cache, need to
use the current user as part of the connection cache key. That way
connections for separate users are distinct, and we always use one
with the correct authorization.
3) The task tracker needs to keep track of the username associated with
a task, so it can use it when establishing connections outside the
originating session.
This commit adds a fast shard pruning path for INSERTs on
hash-partitioned tables. The rationale behind this change is
that if there exists a sorted shard interval array, a single
index lookup on the array allows us to find the corresponding
shard interval. As mentioned above, we need a sorted
(wrt shardminvalue) shard interval array. Thus, this commit
updates shardIntervalArray to sortedShardIntervalArray in the
metadata cache. Then uses the low-level API that is defined in
multi_copy to handle the fast shard pruning.
The performance impact of this change is more apparent as more
shards exist for a distributed table. Previous implementation
was relying on linear search through the shard intervals. However,
this commit relies on constant lookup time on shard interval
array. Thus, the shard pruning becomes less dependent on the
shard count.
Hadi Moshayedi
Hadi Moshayedi
Marco Slot
Jason Petersen
Onder Kalaci
velioglu
mehmet furkan şahin
Murat Tuncer
Brian Cloutier
Metin Doslu
Murat Tuncer
Brian Cloutier
Jason Petersen
Marco Slot
Burak Yucesoy
Andres Freund
Eren Basak
Eren
Matthew Seaman