Various packages within Volt Active Data have been updated to eliminate known security vulnerabilities, including:

The following limitations in previous versions have been resolved.

VoltDB does not use any of the components implicated in the published CVEs related to Log4J. However, to avoid any confusion, VoltDB has replaced the Log4J library with reload4J, a drop-in replacement that replicates the log4J namespace and functionality, but eliminates all known security vulnerabilities.

The goal of the @Statistics system procedure's INITIATOR selector is to help users understand and visualize the performance of their application's queries and transactions. However, in the past, these statistics inadvertently included data associated with import processing. To make the output of the INITIATOR statistics more accurate, import processes are no longer included in the procedure results.

The following limitations in previous versions have been resolved.

The following limitations in previous versions have been resolved.

A number of functional improvements have been made to the VoltDB Management Center (VMC), including:

The following limitations in previous versions have been resolved.

The following limitations in previous versions have been resolved.

VoltDB constrains the size of messages sent between cluster nodes and will cancel transactions that exceed the limit. However, in rare situations, the system itself can generate overly large messages and cause a "bad message length" error. This release adds additional hexadecimal information to the logs when this happens, to help identify the root cause of the error.

The following limitations in previous versions have been resolved.

The jQuery libraries used by the VoltDB Management Center have been updated to the following versions to address security vulnerabilities:

The following limitation in previous versions has been resolved.

The LIMIT PARTITION ROWS feature was deprecated in Version 9 and will be removed in Version 11. This is a change to the VoltDB schema syntax that is not forward compatible.

This means that if your database schema still contains the LIMIT PARTITION ROWS syntax, you need to remove the offending clause before upgrading to the upcoming major release. Fortunately, there is a simple process for doing this. You can use the ALTER TABLE {table-name} DROP LIMIT PARTITION ROWS statement to correct the table schema while the database is running and with no impact to the database contents.

This release includes a number of improvements to the licensing and management of VoltDB software. These improvements include:

The following limitations in previous versions have been resolved.

The following limitations in previous versions have been resolved:

The following limitations in previous versions have been resolved:

The following limitations in previous versions have been resolved:

The following limitations in previous versions have been resolved:

The voltadmin DR DROP command removes a cluster from a cross datacenter replication conversation. Previously, this procedure could take up to 90 seconds. The command has been optimized to significantly reduce the time it takes to complete.

The following limitations in previous versions have been resolved:

VoltDB has completed testing and qualification of the Red Hat Enterprise and CentOS version 8 operating systems as supported platforms for running VoltDB in development and production.

The following limitations in previous versions have been resolved:

Save and Restore:

Monitoring:

Export:

Other:

The export subsystem has been rewritten to provide significant improvements in both performance and reliability, as well as accommodate planned future enhancements. The new subsystem is available to all customers using the VoltDB Enterprise Edition. The key advantages of the new export subsystem are:

VoltDB 9.2 introduced tasks, which let you schedule stored procedures for execution on a repeating schedule. Tasks are now complete and ready for production use. In addition, this release extends support to include custom tasks, where you can dynamically adjust the procedure called, the parameters to that procedure, the interval between calls, or any combination of the three based on the results of the previous invocation. You write custom tasks as Java classes that set the attributes of the next task run and identify a callback method to invoke once the procedure completes. See the chapter on custom tasks in the VoltDB Guide to Performance and Customization for details. There is also a sample custom task available in the VoltDB github repository.

The VoltDB Enterprise Edition now lets you control the thread pools used to execute the export subsystem. A thread pool defines the number of threads used to run processes concurrently. The more threads available, the more concurrent processes and therefore the more throughput. However, more threads means more system resources are consumed. Thread pools let you tune the export subsystem to balance throughput and resource utilization against your application requirements. You specify the thread pools in the database configuration file.

You can define a default thread pool for export connectors, using the defaultpoolsize attribute to the <export> element. You can also assign specific pool sizes to individual export connectors by defining a named thread pool in the <threadpools> element and assigning it to a connector using the threadpool attribute of the <configuration> element. For example:

<export defaultpoolsize="2">
    <configuration target="log" type="file" threadpool="logpool">
        <property name="type">csv</property>
        <property name="nonce">Myapp</property>
    </configuration>
</export>

<threadpools>
    <pool name="logpool" size="5"/>
</threadpools>

When using K-safety, it is important that stored procedures produce deterministic results so multiple copies of a partition can run transactions concurrently with predictable results. Without consistent results in every partition, the data could diverge. To avoid this, traditionally VoltDB would stop the database if a non-deterministic result was detected (referred to as a "hash mismatch" in the log file and error messages).

Writing deterministic stored procedures is still important. But VoltDB now takes less disruptive action when divergence is detected. Rather than stopping the database every time a mismatch is detected, in most cases VoltDB now shuts down the extra copies of the partitions and runs in a single-copy reduced mode, eliminating the possibility of divergence. Of course, in the reduced mode, the cluster is no longer K-safe, which should be corrected as soon as possible. But until the offending procedure can be replaced and the cluster restarted, the database will continue to run and continue to process transactions.

VoltDB easily handles databases with large numbers of tables. However, the schema also contains any stored procedures and auxiliary class JAR files loaded into the database. For databases using extremely large JAR files — most noticeably databases including machine learning (ML) models — it was possible to exceed the database's 50MB limit for the schema.

VoltDB has now been rewritten to accommodate schema of arbitrary size. Note, however, that limitations on the size of individual tables (that is, no more than 1,024 columns per table and not to exceed 2MB for each table) are still in effect.

In addition to the reliability tests that are run on an ongoing basis, Java 11 was run through a series of performance tests to assess its impact on VoltDB, with a particular focus on the different garbage collection schemes. It turns out that all of the garbage collectors available in Java 11 perform as well or better than previous versions, in several tests notably reducing the size and frequency of GC pauses.

Schema changes (and configuration updates) must be treated as multi-partition transactions in VoltDB. As a result, frequent schema changes can impact ongoing throughput and latency. This is particularly noticeable for large schema. Version 9.3 provides performance improvements to reduce the time required to apply schema changes and, as a result, reduce the impact on application latency.

The current release improves the content and structure of VoltDB performance and management data provided to the New Relic monitoring suite.

All SNMP traps are now recorded in the log file. Previously, SNMP traps were sent, but not recorded. Log entries provide additional confirmation that SNMP events are triggered.

The LIMIT PARTITION ROWS clause of the CREATE TABLE statement is being deprecated. LIMIT PARTITION ROWS was designed to avoid data overflowing the available space in the database. However, it was an all-or-nothing setting: once you exceeded the limit the database was automatically paused. Transactions could not continue until you manually reduced the row count. The feature itself did nothing to resolve the situation.

LIMIT PARTITION ROWS is being deprecated and replaced by two significantly better capabilities now available in VoltDB. USING TTL lets you automatically and incrementally purge old data from tables and scheduled tasks let you build complex responsive algorithms for monitoring, managing, and resolving data volumes.

Although still available in V9.3, customers should replace any use of the LIMIT PARTITION ROWS clause at their earliest convenience, because the syntax will be removed from the product in a future major release.

In VoltDB V8.0, the interface for writing custom export clients was updated and replaced with a new API. To accommodate old clients, the export system uses the method signatures to distinguish between the old and new interface. However, the old interface is now being deprecated and support will be removed in the next major release.

Any custom clients that use the old signature, where the onBlockStart() andonBlockCompletion() accept no arguments and processRow() expects two, should be updated to use the new interface. See the chapter on custom export and import in the VoltDB Guide to Performance and Customization for information on the new interface.

The export connector for RabbitMQ is now deprecated and will be removed in a future release.

The following libraries used by VoltDB have been updated to ensure the latest security and performance patches are applied:

In addition to the new features and capabilities described above, the following limitations in previous versions have been resolved:

The packaging of VoltDB in the Maven Central Repository has been adjusted to ensure the correct artifacts are provided.

The handling of secure export connections (using SSL/TLS) through the JDBC interface has been improved. Specifically, the requirement for a truststore when using a commercial certificate has been removed.

The Kafka acks property determines whether VoltDB waits for acknowledgement of receipt from the Kafka brokers. Previously, the default was set to "1" (one), but the recommendation was to set it to "all" to protect against the loss of records if the Kafka brokers fail. The default has changed to match the recommended setting.

Existing customers who use Kafka export but do not explicitly set the acks property may notice a slight change in export latency. The new default of "all" is the recommended setting. However, if you are willing to accept less durability on the part of the Kafka brokers, you can explicitly set the property back to "1" to replicate previous behavior.

VoltDB 9.1 introduced two new beta features: the ability to export data directly from tables using the CREATE TABLE... EXPORT TO TARGET statement and the MIGRATE statement to simplify the export and deletion of records. These features are now fully supported for production use.

VoltDB 9.2 introduces a new feature, scheduled tasks. Tasks let you schedule the repeated execution of stored procedures at a set interval or using a cron-style declaration. You schedule tasks using the CREATE TASK statement. There is also a corresponding @Statistics selector, TASK, and ALTER TASK and DROP TASK statements. See the section on scheduling tasks or the reference pages describing each statement in the Using VoltDB manual for details.

One aspect of scheduling tasks is defining how they run; that is, as a single multi-partition transaction or as separate transactions on each partition. To support the latter execution model, a new type of stored procedure is being introduced, the directed procedure. Directed procedures are partitioned procedures in that each instance of the procedure runs on a separate partition. However, directed procedures do not have a specific partitioning value. You use the CREATE TASK statement or the Java callAllPartitionProcedure method to have a separate instance of the procedure run on every partition of the database. See the section on directed procedures in the Using VoltDB manual for details.

The VoltDB Management Center (VMC) has a new tab, Export, which provides enhanced statistics and graphs that help analyze the performance of export connectors.

The CREATE FUNCTION statement lets you declare a user-defined scalar function. Starting with VoltDB 9.2, you can develop and declare user-defined aggregate functions as well using the CREATE AGGREGATE FUNCTION statement. Aggregate functions process data from multiple rows and return a single aggregated value. See the chapter on user-defined functions in the VoltDB Guide to Performance and Customization for details.

You can now query statistics from the VoltDB @Statistics system procedure as if the results were SQL tables. You can use the querystats directive in sqlcmd or the new @QueryStats system procedure specifying the column names from the @Statistics results in the selection expression. For example, the following sqlcmd command aggregates the row count from all of the partitions for each table using the TABLE selector:

$ sqlcmd 
1> querystats select table_name, sum(tuple_count) from statistics(table,0) group by table_name;

Note that not all SQL syntax is supported as input to the querystats parser and for the initial release any syntax errors are reported in the server log but not reported to the user's console. See the descriptions of sqlcmd or the @QueryStats system procedure in the Using VoltDB manual for further details.

You can now control how frequently the buffers for database replication (DR) and export are flushed. Normally, DR and export data is buffered until a certain amount of data is ready and then the data is sent as a batch. To avoid small amounts of data lingering in the buffer, there is a time limit after which, the data is sent even if the full batch size has not been reached. This time limit is called the "flush interval". You can now control these settings in the configuration file by setting a system-wide minimum and separate flush intervals for DR and export. See the VoltDB Administrator's Guide for details on configuring flush intervals.

In conjunction with the new configuration options for flushing buffers, an older undocumented attribute for setting the DR flush interval is being deprecated. The flushInterval attribute of the <dr> element is deprecated and will be removed in a future major release. In the meantime, this attribute will continue to operate and will supersede the new settings and defaults, so as not to change existing behavior for any customers who may have used this feature. However, those users are strongly encouraged to switch to the new, supported feature at their earliest convenience to avoid problems when the deprecated attribute is removed from the product in the future.

In addition to the new features and capabilities described above, the following limitations in previous versions have been resolved:

The license for the VoltDB JAR file in Maven has been changed from AGPL to an MIT license.

Previously, you could expand a running cluster to increase capacity by starting the new server with the voltdb start --add command. However, until now there was no way to reduce the cluster size without stopping and reconfiguring the database. With the introduction of the voltadmin resize command, you can now elastically shrink a running cluster as well. The voltadmin resize command tests the cluster to make sure it can be reduced in size, tells you which nodes will be removed, and then starts the resize process. See the section on "Removing Nodes with Elastic Scaling" in the Using VoltDB manual for details on reducing the size of a running VoltDB cluster.

VoltDB V9.0 introduced migration as a beta feature, making it possible to automate the data lifecycle by migrating data to an external resource before deleting it from the VoltDB database. With V9.1, the migration feature is now fully integrated, tested, and extended to round out its capabilities.

See the description of the CREATE TABLE statement in the Using VoltDB manual for details on automating the migration of data to external targets.

Using the MIGRATE TO TARGET clause with USING TTL automates the export of data to an external target before the data is deleted. In addition, you can now use the MIGRATE TO TARGET clause by itself — without the USING TTL clause — for situations where you want to manually control when the data is migrated. To do this, you can use the new MIGRATE statement to manually initiate the migration during a transaction. Note that you can also use the MIGRATE to statement for tables defined with both MIGRATE TO TARGET and USING TTL, in which case an explicit MIGRATE statement can preemptively migrate the data before the TTL value is reached. See the description of the MIGRATE statement in the Using VoltDB manual for details.

VoltDB 9.1 introduces a new feature that makes it possible to connect tables (not just streams) to export targets. If a table is declared with the EXPORT TO TARGET clause, just like a stream, any data inserted into the table is passed to the export connector. This simplifies applications that want to stream incoming data to external systems, where previously there had to be both a table and a matching stream.

For example, the following table declaration allows any data inserted into the Alerts table to also be exported to the Messagelog target:

CREATE TABLE Alerts 
  EXPORT TO TARGET Messagelog 15519
   ( eventtime TIMESTAMP,
     sender VARCHAR(16),
     message VARCHAR(128)
   ); 

You can also customize which events trigger export. Triggering events include inserts, updates (both before and after the update) and deletes. By default, only inserts trigger export; you can specify a different list of triggers with the ON clause:

CREATE TABLE user 
  EXPORT TO TARGET userarchive
    ON INSERT, UPDATE, DELETE 
   ( username VARCHAR(128),
     id BIGINT,
     lastlogin TIMESTAMP
   ); 

See the description of the CREATE TABLE statement in the Using VoltDB manual for more information about associating tables with export targets.

VoltDB limits each transaction to 50 MB of results. In fact, each transaction fragment is limited to 50MB. However, for multi-partition transactions, this means each partition can return up to 50MB of data to the coordinator. For large clusters with many unique partitions, it is possible for all this data to exceed the allocated Java heap for the coordinator, causing that node to fail with an out of memory error.

To avoid this situation, VoltDB now limits the amount of data each fragment can return, based on the current maximum heap size and number of partitions. By default, each partition in a multi-partition transaction is only allowed to return the lesser of 65% of the maximum heap size divided by the number of unique partitions (sitesperhost * number of nodes / k+1 ) or 50MB. The limit is further reduced for read-only multi-partition transactions to accommodate for the fact that multiple read-only transactions can be run simultaneously. If, at runtime, a fragment exceeds the limit, it throws an exception and the transaction rolls back gracefully.

You can adjust this per-partition multi-partition response limit by setting the environment variable MP_MAX_TOTAL_RESP_SIZE. You can either set it as the percentage of max heap to use in the calculation (by using the percent sign) or as a specific number of bytes (by using an integer value with no suffix). For example, to allow a maximum of only 50% of the allowable heap size, you can set the following environment variable before starting the server:

$ export MP_MAX_TOTAL_RESP_SIZE="50%"
$ voltdb start -D ~/mydb ...

You can set MP_MAX_TOTAL_RESP_SIZE as an environment variable or as a Java system property through VOLTDB_OPTS:

$ export VOLTDB_OPTS="-DMP_MAX_TOTAL_RESP_SIZE=50%"
$ voltdb start -D ~/mydb ...

There is a new command, voltadmin show license, that lists information about the current license in use by a running VoltDB cluster. You can get similar information programmatically using the @SystemInformation system procedure with the LICENSE selector.

In addition to the new features and capabilities described above, the following limitations in previous versions have been resolved:

The requirements on the underlying operating system and software environment have been updated for VoltDB V9.0. The older CentOS and RHEL version 6.6 is no longer supported and Java 11 support has been added. In addition, support for kafka 0.8.2 has been dropped and Kafka import and export now requires version 0.10.2 or later. See the VoltDB Administrator's Guide for details on the platform requirements for running VoltDB clusters.

VoltDB now supports both Java 8 and Java 11.

VoltDB 8.4 introduced a new feature, USING TTL ("time to live"), that lets you define when records expire and can be deleted. This feature simplifies application design by automatically removing old data from the database based on settings you define in the table schema. With VoltDB 9.0, this feature is extended to include the migration of deleted data to other systems for archival purposes, as described next.

The code that supports export of data to external systems has been rewritten to provide flexibility, improve reliability, reduce system resource utilization, and support new and future product features. The new export system reinforces the durability of data queued to the export connectors across unexpected system and network failures and allows export to be extended to add new capabilities.

The first two new capabilities are:

Export now starts when the stream is defined, not when the target is defined. Previously, stream data was not queued for export until a valid export connector was configured and connected. Starting with VoltDB 9.0, data written to streams declared with the EXPORT TO TARGET clause are queued for export whether the target is configured or not. Similarly, the queued data is removed as soon as the stream itself is removed with the DROP STREAM statement.

Also, export is now an enterprise feature. The VoltDB Community Edition provides access to two streams per database, so users have access to basic export functionality. But for unlimited access to export and migration features, the Enterprise Edition is required.

Previously, database replication (DR) required the schema of the cooperating databases to match for all DR tables. So updating the schema required a pause while all of the affected databases were updated. Starting with 9.0, this limitation has been loosened. DR continues even if the schema are different. So it is possible to update the schema without interrupting ongoing transactions.

Of course, it is not possible for VoltDB to resolve individual transactions if the schema differ. So if a DR consumer (either a replica in passive DR or an XDCR cluster in active replication) receives a binary log where the schema of the affected table(s) does not match, DR will stall and wait for the schema to be updated to match the incoming data. Therefore, care must be taken when updating the schema to ensure that no transactions that are affected by the schema change are processed during the interval when the clusters' schema do not match. See the sections on updating DR schema for passive and active DR for more information.

A new version of the VoltDB JSON API, 2.0, is now available. The original JSON interface provides complete information about the schema for the data being returned, including separate entries for the data, the column names, and datatypes. The 2.0 API returns a much more compact result set with each row represented by an associative array with each element consisting of the column name and value.

There was an undocumented feature of the @Statistics system procedure that reported on how busy the execution queues for the individual partitions are. This data is now supported as the IDLETIME selector. See the description of the @Statistics system procedure in the Using VoltDB manual for details.

The I/O activity that JVM stats generates can interface with performance for applications like VoltDB, to the point where it can cause cluster nodes to disconnect. VoltDB now automatically disables JVM stats if /tmp is not defined as tmpfs (temporary in-memory storage).

Export streams (that is streams defined with the EXPORT TO TARGET clause) are no longer reported under the TABLE statistics of the @Statistics system procedure. Export streams are now reported under the EXPORT selector and tables and streams without export are reported under TABLE.

The voltdb init command now allows you to specify multiple files as arguments to the --schema and --classes flags. Separate multiple files with commas. You can also use the asterisk (*) as a wildcard character. For example, the following command initializes a root directory with two schema files, plus all the JAR files from one folder and another JAR file from the current working directory:

$ voltdb init -D ~/mydb \
     --schema=customers.sql,companies.sql \
     --classes=storedprocs/*.jar,myfunctions.jar

It is also possible to specify multiple schema and classes files when configuring VoltDB for use in Kubernetes. See in the readme file in the tools/kubernetes/ subfolder where VoltDB is installed for details.

The Log4J logger for elastic operations such as adding new cluster nodes on the fly has been renamed from JOIN to ELASTIC

The following change has been made to improve security and eliminate potential threats:

In addition to the new features and capabilities described above, the following limitations in previous versions have been resolved:

VoltDB reserves the subfolder "segments" under the command log directory for storing the actual command log files. Do not add, remove, or modify any files in this directory. In particular, do not set the command log snapshot directory to a subfolder "segments" of the command log directory, or else the server will hang on startup.

Because DR data is buffered on the master database and then delivered asynchronously to the replica, there is always the danger that data does not reach the replica if a master node stops. This situation is mitigated in a K-safe environment by all copies of a partition buffering on the master cluster. Then if a sending node goes down, another node on the master database can take over sending logs to the replica. However, if multiple nodes go down and rejoin in rapid succession, it is possible that some buffered DR data — from transactions when one or more nodes were down — could be lost when another node with the last copy of that buffer also goes down.

If this occurs and the replica recognizes that some binary logs are missing, DR stops and must be restarted.

To avoid this situation, especially when cycling through nodes for maintenance purposes, the key is to ensure that all buffered DR data is transmitted before stopping the next node in the cycle. You can do this using the @Statistics system procedure to make sure the last ACKed timestamp (using @Statistitcs DR on the master cluster) is later than the timestamp when the previous node completed its rejoin operation.

Bulk operations, such as large deletes, inserts, or updates are possible within a single stored procedure. However, if the binary logs generated for DR are larger than 45MB, the operation will fail. To avoid this situation, it is best to break up large bulk operations into multiple, smaller transactions. A general rule of thumb is to multiply the size of the table schema by the number of affected rows. For deletes and inserts, this value should be under 45MB to avoid exceeding the DR binary log size limit. For updates, this number should be under 22.5MB (because the binary log contains both the starting and ending row values for updates).

There are a number of VoltDB features that help manage the database by constraining memory size and resource utilization. These features are extremely useful in avoiding crashes as a result of unexpected or unconstrained growth. However, these features could interfere with the normal operation of DR when passing data from one cluster to another, especially if the two clusters are different sizes. Therefore, as a general rule of thumb, DR overrides these features in favor of maintaining synchronization between the two clusters.

Specifically, DR ignores any resource monitor limits defined in the deployment file when applying binary logs on the consumer cluster. DR also ignores any partition row limits defined in the database schema when applying binary logs. This means, for example, if the replica database in passive DR has less memory or fewer unique partitions than the master, it is possible that applying binary logs of transactions that succeeded on the master could cause the replica to run out of memory. Note that these resource monitor and tables row limits are applied on any original transactions local to the cluster (for example, transactions on the master database in passive DR).

Database Replication (DR) now supports replication across clusters of different sizes. However, if the replica cluster is smaller than the master cluster, it may require a significantly larger Java heap setting. Specifically, if the replica has fewer unique partitions than the master, each partition on the replica must manage the incoming binary logs from more partitions on the master, which places additional pressure on the Java heap.

A simple rule of thumb is that the worst case scenario could require an additional P * R * 20MB space in the Java heap , where P is the number of sites per host on the replica server and R is the ratio of unique partitions on the master to partitions on the replica. For example, if the master cluster is 5 nodes with 10 sites per host and a K factor of 1 (i.e. 25 unique partitions) and the replica cluster is 3 nodes with 8 sites per host and a K factor of 1 (12 unique partitions), the Java heap on the replica cluster may require approximately 320MB of additional space in the heap:

An alternative is to reduce the size of the DR buffers on the master cluster by setting the DR_MEM_LIMIT Java property. For example, you can reduce the DR buffer size from the default 10MB to 5MB using the VOLTDB_OPTS environment variable before starting the master cluster.

Changing the DR buffer limit on the master from 10MB to 5MB proportionally reduces the additional heap size needed. So in the previous example, the additional heap on the replica is reduced from 320MB to 160MB.

The voltadmin status --dr command provides real-time status on the state of database replication (DR). Normally, this includes the status of the current cluster as well as other clusters in the DR environment. (For example, both the master and replica in passive DR or all clusters in XDCR.) However, if the clusters are configured to use different port numbers for the client port, VoltDB cannot reach the other clusters and the command hangs until it times out waiting for a response from the other clusters.

Part of the replication process for XDCR is to verify that the record's starting and ending states match on both clusters, otherwise known as conflict resolution. To do that, XDCR must find the record first. Finding uniquely indexed records is efficient; finding non-unique records is not and can impact overall database performance.

To make you aware of possible performance impact, VoltDB issues a warning if you declare a table as a DR table and it does not have a unique index.

You cannot start XDCR if both databases already have data in the DR tables. Only one of the two participating databases can have preexisting data when DR starts for the first time.

When starting XDCR for the first time, where one database already contains data, a snapshot of that data is sent to the other database. While receiving and processing that snapshot, the receiving database is paused. That is, it is in read-only mode. Once the snapshot is completed and the two database are synchronized, the receiving database is automatically unpaused, resuming normal read/write operations.

Normally, XDCR will automatically restart where it left off after one of the clusters stops and recovers from its command logs (using the voltdb recover command). However, if the workload is predominantly multi-partition write transactions, a failed cluster may not be able to restart XDCR after it recovers. In this case, XDCR must be restarted from scratch, using the content from one of the clusters as the source for synchronizing and recreating the other cluster (using the voltdb create --force command) without any content in the DR tables.

TRUNCATE TABLE is optimized to delete all data from a table rather than deleting tuples row by row. This means that the binary log does not identify which rows are deleted. As a consequence, a TRUNCATE TABLE statement and a simultaneous write operation to the same table can produce a conflict that the XDCR clusters cannot detect or report in the conflict log.

Therefore, do not use TRUNCATE TABLE with XDCR. Instead, explicitly delete all rows with a DELETE statement and a filter. For example, DELETE * FROM table WHERE column=column ensures all deleted rows are identified in the binary log and any conflicts are accurately reported. Note that DELETE FROM table without a WHERE clause is not sufficient, since its execution plan is optimized to equate to TRUNCATE TABLE.

It is possible to place a limit on the number of rows that any partition can hold for a specific table using the LIMIT PARTITION ROWS clause of the CREATE TABLE statement. When close to the limit, transactions on either or both clusters can exceed the limit simultaneously, resulting in a potentially large number of delete operations that then generate conflicts when the the associated binary log reaches the other cluster.

VoltDB provides a way to generate a deterministically unique ID within a stored procedure using the getUniqueId method. This method guarantees uniqueness within the current cluster. However, the method could generate the same ID on two distinct database clusters. Consequently, when using XDCR, you should combine the return values of VoltProcedure.getUniqueId with VoltProcedure.getClusterId, which returns the current cluster's unique DR ID, to generate IDs that are unique across all clusters in your environment.

TTL, or time to live, is a feature that automatically deletes old records based on a timestamp or integer column. For replicated tables, the process of checking whether records need to be deleted is performed as a write transaction — even if no rows are deleted. As a consequence, any replicated DR table with TTL defined will generate frequent DR log entries, whether there are any changes or not, significantly increasing DR traffic.

Because of the possible performance impact this behavior can have on the database, use of TTL with replicated tables and DR is not recommended at this time.

A bug in the Apache Kafka client can result in file descriptors being allocated but not released if the producer.type attribute is set to "sync" (which is the default). The consequence is that the system eventually runs out of file descriptors and the VoltDB server process will crash.

Until this bug is fixed, use of synchronous Kafka export is not recommended. The workaround is to set the Kafka producer.type attribute to "async" using the VoltDB export properties.

If, while Kafka import is enabled, the Kafka broker that VoltDB is connected to stops (for example, if the server crashes or is taken down for maintenance), some messages may be lost between Kafka and VoltDB. To ensure no data is lost, we recommend you disable VoltDB import before taking down the associated Kafka broker. You can then re-enable import after the Kafka broker comes back online.

There is an issue with the Kafka importer where, if multiple nodes in the cluster fail and restart, the importer can lose track of some of the data that was being processed when the nodes failed. Normally, these pending imports are replayed properly on restart. But if multiple nodes fail, it is possible for some in-flight imports to get lost. This issue will be addressed in an upcoming release.

When entering a multi-line statement at the sqlcmd prompt, if a line ends in a comment (indicated by two hyphens) and the comment contains an unmatched single quote character, the following lines of input are not interpreted correctly. Specifically, the comment is incorrectly interpreted as continuing until the next single quote character or a closing semi-colon is read. This is most likely to happen when reading in a schema file containing comments. This issue is specific to the sqlcmd utility.

A fix for this condition is planned for an upcoming point release

VoltDB currently intercepts assertions as part of its handling of stored procedures. Attempts to use assertions in stored procedures for debugging or to find programmatic errors will not work as expected.

The UPPER() and LOWER() functions return a string converted to all uppercase or all lowercase letters, respectively. However, for the initial release, these functions only operate on characters in the ASCII character set. Other case-sensitive UTF-8 characters in the string are returned unchanged. Support for all case-sensitive UTF-8 characters will be included in a future release.

There is a problem with how the math library used to build the C++ client library handles large decimal values on 32-bit operating systems. As a result, the C++ library cannot serialize and pass Decimal datatypes reliably on these systems.

Note that the C++ client interface can send and receive Decimal values properly on 64-bit platforms.

Enabling SNMP can take up to 2 minutes to complete. This delay does not always occur and can vary in length. If SNMP is enabled when the database server starts, the delay occurs after the server logs the message "Initializing SNMP" and before it attempts to connect to the cluster. If you enable SNMP while the database is running, the delay can occur when you issue the voltadmin update command or modify the setting in the VoltDB Management Center Admin tab. This issue results from a Java constraint related to secure random numbers used by the SNMP library.

With VoltDB V7.0, cross datacenter replication (XDCR) supports multiple clusters in an XDCR network. However, the VoltDB Management Center currently reports on only one such connection per cluster. In the future, the Management Center will provide monitoring and statistics for all connections to the current cluster.

Any time the database schema or stored procedures are changed, the data table showing stored procedure statistics at the bottom of the Monitor section of the VoltDB Management Center get reset. As soon as new invocations of the stored procedures occur, the statistics table will show new values based on performance after the schema update. Until invocations occur, the procedure table is blank.

The ASSUMEUNIQUE attribute is designed for identifying columns in partitioned tables where the column values are known to be unique but the table is not partitioned on that column, so VoltDB cannot verify complete uniqueness across the database. Using interactive DDL, you can create a table with a column marked as ASSUMEUNIQUE, but if you try to partition the table on the ASSUMEUNIQUE column, you receive an error. The solution is to drop and add the column using the UNIQUE attribute instead of ASSUMEUNIQUE.

You cannot add or remove a column constraint such as UNIQUE or ASSUMEUNIQUE using the ALTER TABLE ALTER COLUMN statement. Instead to add or remove such constraints, you must first drop then add the modified column. For example:

ALTER TABLE employee DROP COLUMN empID;
ALTER TABLE employee ADD COLUMN empID INTEGER UNIQUE;

For partitioned tables, the value of the column used to partition the table determines what partition the row belongs to. If you use UPDATE to change this value and the new value belongs in a different partition, the UPDATE request will fail and the stored procedure will be rolled back.

Updating the partition column value may or may not cause the record to be repartitioned (depending on the old and new values). However, since you cannot determine if the update will succeed or fail, you should not use UPDATE to change the value of partitioning columns.

The workaround, if you must change the value of the partitioning column, is to use both a DELETE and an INSERT statement to explicitly remove and then re-insert the desired rows.

Starting with VoltDB 6.0, ambiguous column references are no longer allowed. For example, if both the Customer and Placedorder tables have a column named Address, the reference to Address in the following SELECT statement is ambiguous:

SELECT OrderNumber, Address FROM Customer, Placedorder
   . . .

Previously, VoltDB would select the column from the leftmost table (Customer, in this case). Ambiguous column references are no longer allowed and you must use table prefixes to disambiguate identical column names. For example, specifying the column in the preceding statement as Customer.Address.

A corollary to this change is that a column declared in a USING clause can now be referenced using a prefix. For example, the following statement uses the prefix Customer.Address to disambiguate the column selection from a possibly similarly named column belonging to the Supplier table:

SELECT OrderNumber, Vendor, Customer.Address
   FROM Customer, Placedorder Using (Address), Supplier
    . . .

VoltDB opens a file descriptor for every client connection to the database. In normal operation, this use of file descriptors is transparent to the user. However, if there are an inordinate number of concurrent client connections, or clients open and close many connections in rapid succession, it is possible for VoltDB to exceed the process limit on file descriptors. When this happens, new connections may be rejected or other disk-based activities (such as snapshotting) may be disrupted.

In environments where there are likely to be an extremely large number of connections, you should consider increasing the operating system's per-process limit on file descriptors.

There are two ways to access additional resources in a VoltDB database. You can place the resources in the /lib folder where VoltDB is installed on each server in the cluster or you can include the resource in a subfolder of a JAR file you add using the sqlcmd LOAD CLASSES directive. Adding resources via the /lib directory is useful for stable resources (such as third-party software libraries) that do not require updating. Including resources (such as XML files) in the JAR file is useful for resources that may need to be updated, as a single transaction, while the database is running.

LOAD CLASSES is used primarily to load classes associated with stored procedures and user-defined functions. However, it will also load any additional resource files included in subfolders of the JAR file. You can remove classes that are no longer needed using the REMOVE CLASSES directive. However, there is no explicit command for removing other resources.

Consequently, if you rename resources or move them to a different location and reload the JAR file, the database will end up having multiple copies. Over time, this could result in more and more unnecessary memory being used by the database. To remove obsolete resources, you must first reinitialize the database root directory, start a fresh database, reload the schema (including the new JAR files with only the needed resources) and then restore the data from a snapshot.

If a server has multiple network interfaces (and therefore multiple IP addresses) VoltDB will, by default, open ports on all available interfaces. You can limit the ports to an single interface in two ways:

Also, when using an IP address to reference a server with multiple interfaces in command line utilities (such as voltadmin stop node), use the @SystemInformation system procedure to determine which IP address VoltDB has selected to identify the server. Otherwise, if you choose the wrong IP address, the command might fail.