gbase数据、南大通用产品文档：GBase8agbase_kafka_enable_transaction

取值：[0|1]
默认值：0
说明：加载和导出到kafka 时，设置是否开启kafka 事务性功能，默认关闭。
值为1代表开启事务性功能，
开启后可以保证所有导出的消息属于同一事务，
即其中一条消息产生失败，整个事务提交失败。同样，开启事务性功能后，
Kafka 数据导入到8a 集群也只有提交成功的事务内的消息可以被导入。
修改方式：
可使用set 语句修改值也可在配置文件中修改值。
适用于session、
global
范围均可。

High-Performance Scalability
GBase 8a MPP Cluster can expand its computing and storage capacity by adding
server nodes, supporting clusters with hundreds or thousands of nodes, and the
expansion process can be flexibly monitored and managed with features such as
monitoring, pausing, resuming, and canceling. The system supports online
expansion, with linear performance improvement, without interrupting the current
system operation, and supports adding multiple nodes at once. It also supports
flexible expansion at different levels such as instance-level, database-level, and
table-level.

GBase 8a MPP Cluster Technical White Paper
- 16 -
General Data Technology Co., Ltd.
GBase 8a MPP Cluster has high-performance scaling capabilities:
1)
Online dynamic expansion of cluster nodes;
2)
Each node can process 100TB of effective data while providing both
computing and storage capabilities;
3)
Execution
scheduling
nodes
and
data
computing
nodes
can
be
independently scaled as needed.
Figure 3-5. GBase 8a MPP ClusterExpansion Technology Diagram.
The last "Execution and Scheduling Node" and "Data Storage and Computation
Node" in the above figure are newly expanded nodes based on the original nodes.
The scheduling nodes or storage and computation nodes can be expanded. Because
GBase 8a MPP Cluster adopts a high-performance single-node MPP architecture, it
can ensure smooth expansion and linear performance growth during cluster
expansion, as shown in the following figure:

GBase 8a MPP Cluster Technical White Paper
General Data Technology Co., Ltd.
- 17 -
Figure 3-6. GBase 8a MPP Cluster high-performance and high-scalability diagram.

空的临时 dbspace 归还空间到存储池
（SQL 管

GBase 8s 管理员参考
南大通用数据技术股份有限公司 - 732 -

理 API）
随同 admin() 或 task() 函数，使用 drop tempdbspace to storagepool 参数来从空的
临时 dbspace 归还空间到存储池。
语法

元素
描述
关键考虑
tempdbspace
空的临时 dbspace 名称。

示例
下列命令删除名为 tempdbs1 的临时 dbspace，并将所有释放的空间添加到存储池。
EXECUTE FUNCTION task("drop tempdbspace to storagepool", "tempdbs1");