FLINKCDC连接oracle导致归档日志暴增

devtools/2024/12/22 3:02:01/

前言

前段时间再用flinkcdc连接oracle的时候，oracle的归档日志疯狂的飙升,我经常去到归档目录下查看占用的内存情况。。

情况

在使用flinkcdc连接oracle的时候，发现归档日志飙升，查看了很多文档，该配置了都配置了，

# 连接状态的时候，回一直飙升
root@1b612c91b33b:/u01/app/oracle/archive# du -h
674M    .

flinkcdc程序

        Properties properties = new Properties();properties.setProperty("debezium.database.tablename.case.insensitive", "false");properties.setProperty("debezium.log.mining.strategy", "online_catalog");
//        properties.setProperty("database.tablename.case.insensitive", "false");
//        properties.setProperty("log.mining.strategy", "online_catalog");//持续挖掘日志properties.setProperty("debezium.log.mining.continuous.mine", "true");SourceFunction<String> sourceFunction = OracleSource.<String>builder().hostname("192.168.1.9").port(1521).database("XE") // monitor XE database.schemaList("flinkuser") // monitor inventory schema.tableList("flinkuser.SYS_USER") // monitor products table.username("flinkuser").password("flinkpw").startupOptions(StartupOptions.initial()).debeziumProperties(properties).deserializer(new JsonDebeziumDeserializationSchema()) // converts SourceRecord to JSON String.build();StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();DataStreamSource<String> sourceStream = env.addSource(sourceFunction).setParallelism(1);// use parallelism 1 for sink to keep message orderingSingleOutputStreamOperator<String> streamOperator = sourceStream.filter(StringUtils::isNotEmpty);streamOperator.setParallelism(1).print();env.execute();

问题解决

问题是怎么解决的呢,将flinkcdc的代码修改一下，，

        properties.setProperty("debezium.database.tablename.case.insensitive", "false");properties.setProperty("debezium.log.mining.strategy", "online_catalog");换成下面两句。。properties.setProperty("database.tablename.case.insensitive", "false");properties.setProperty("log.mining.strategy", "online_catalog");