Paimon Catalog
使用须知
- 数据放在 hdfs 时,需要将 core-site.xml,hdfs-site.xml 和 hive-site.xml 放到 FE 和 BE 的 conf 目录下。优先读取 conf 目录下的 hadoop 配置文件,再读取环境变量
HADOOP_CONF_DIR
的相关配置文件。 - 当前适配的 Paimon 版本为 0.8。
创建 Catalog
Paimon Catalog 当前支持两种类型的 Metastore 创建 Catalog:
- filesystem(默认),同时存储元数据和数据在 filesystem。
- hive metastore,它还将元数据存储在 Hive metastore 中。用户可以直接从 Hive 访问这些表。
基于 FileSystem 创建 Catalog
HDFS
CREATE CATALOG `paimon_hdfs` PROPERTIES (
"type" = "paimon",
"warehouse" = "hdfs://HDFS8000871/user/paimon",
"dfs.nameservices" = "HDFS8000871",
"dfs.ha.namenodes.HDFS8000871" = "nn1,nn2",
"dfs.namenode.rpc-address.HDFS8000871.nn1" = "172.21.0.1:4007",
"dfs.namenode.rpc-address.HDFS8000871.nn2" = "172.21.0.2:4007",
"dfs.client.failover.proxy.provider.HDFS8000871" = "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider",
"hadoop.username" = "hadoop"
);
CREATE CATALOG `paimon_kerberos` PROPERTIES (
'type'='paimon',
"warehouse" = "hdfs://HDFS8000871/user/paimon",
"dfs.nameservices" = "HDFS8000871",
"dfs.ha.namenodes.HDFS8000871" = "nn1,nn2",
"dfs.namenode.rpc-address.HDFS8000871.nn1" = "172.21.0.1:4007",
"dfs.namenode.rpc-address.HDFS8000871.nn2" = "172.21.0.2:4007",
"dfs.client.failover.proxy.provider.HDFS8000871" = "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider",
'hadoop.security.authentication' = 'kerberos',
'hadoop.kerberos.keytab' = '/doris/hdfs.keytab',
'hadoop.kerberos.principal' = 'hdfs@HADOOP.COM'
);
MINIO
CREATE CATALOG `paimon_s3` PROPERTIES (
"type" = "paimon",
"warehouse" = "s3://bucket_name/paimons3",
"s3.endpoint" = "http://<ip>:<port>",
"s3.access_key" = "ak",
"s3.secret_key" = "sk"
);
OBS
CREATE CATALOG `paimon_obs` PROPERTIES (
"type" = "paimon",
"warehouse" = "obs://bucket_name/paimon",
"obs.endpoint"="obs.cn-north-4.myhuaweicloud.com",
"obs.access_key"="ak",
"obs.secret_key"="sk"
);
COS
CREATE CATALOG `paimon_s3` PROPERTIES (
"type" = "paimon",
"warehouse" = "cosn://paimon-1308700295/paimoncos",
"cos.endpoint" = "cos.ap-beijing.myqcloud.com",
"cos.access_key" = "ak",
"cos.secret_key" = "sk"
);
OSS
CREATE CATALOG `paimon_oss` PROPERTIES (
"type" = "paimon",
"warehouse" = "oss://paimon-zd/paimonoss",
"oss.endpoint" = "oss-cn-beijing.aliyuncs.com",
"oss.access_key" = "ak",
"oss.secret_key" = "sk"
);
基于 Hive Metastore 创建 Catalog
CREATE CATALOG `paimon_hms` PROPERTIES (
"type" = "paimon",
"paimon.catalog.type" = "hms",
"warehouse" = "hdfs://HDFS8000871/user/zhangdong/paimon2",
"hive.metastore.uris" = "thrift://172.21.0.44:7004",
"dfs.nameservices" = "HDFS8000871",
"dfs.ha.namenodes.HDFS8000871" = "nn1,nn2",
"dfs.namenode.rpc-address.HDFS8000871.nn1" = "172.21.0.1:4007",
"dfs.namenode.rpc-address.HDFS8000871.nn2" = "172.21.0.2:4007",
"dfs.client.failover.proxy.provider.HDFS8000871" = "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider",
"hadoop.username" = "hadoop"
);
CREATE CATALOG `paimon_kerberos` PROPERTIES (
"type" = "paimon",
"paimon.catalog.type" = "hms",
"warehouse" = "hdfs://HDFS8000871/user/zhangdong/paimon2",
"hive.metastore.uris" = "thrift://172.21.0.44:7004",
"hive.metastore.sasl.enabled" = "true",
"hive.metastore.kerberos.principal" = "hive/xxx@HADOOP.COM",
"dfs.nameservices" = "HDFS8000871",
"dfs.ha.namenodes.HDFS8000871" = "nn1,nn2",
"dfs.namenode.rpc-address.HDFS8000871.nn1" = "172.21.0.1:4007",
"dfs.namenode.rpc-address.HDFS8000871.nn2" = "172.21.0.2:4007",
"dfs.client.failover.proxy.provider.HDFS8000871" = "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider",
"hadoop.security.authentication" = "kerberos",
"hadoop.kerberos.principal" = "hdfs@HADOOP.COM",
"hadoop.kerberos.keytab" = "/doris/hdfs.keytab"
);
基于 Aliyun DLF 创建 Catalog
该功能自 2.1.7 和 3.0.3 版本支持。
CREATE CATALOG `paimon_dlf` PROPERTIES (
"type" = "paimon",
"paimon.catalog.type" = "dlf",
"warehouse" = "oss://xx/yy/",
"dlf.proxy.mode" = "DLF_ONLY",
"dlf.uid" = "xxxxx",
"dlf.region" = "cn-beijing",
"dlf.access_key" = "ak",
"dlf.secret_key" = "sk"
-- "dlf.endpoint" = "dlf.cn-beijing.aliyuncs.com", -- optional
-- "dlf.catalog.id" = "xxxx", -- optional
);
列类型映射
Paimon Data Type | Doris Data Type | Comment |
---|---|---|
BooleanType | Boolean | |
TinyIntType | TinyInt | |
SmallIntType | SmallInt | |
IntType | Int | |
FloatType | Float | |
BigIntType | BigInt | |
DoubleType | Double | |
VarCharType | VarChar | |
CharType | Char | |
VarBinaryType, BinaryType | String | |
DecimalType(precision, scale) | Decimal(precision, scale) | |
TimestampType,LocalZonedTimestampType | DateTime | |
DateType | Date | |
ArrayType | Array | 支持Array嵌套 |
MapType | Map | 支持Map嵌套 |
RowType | Struct | 支持Struct嵌套(2.0.10 和 2.1.3 版本开始支持) |
常见问题
Kerberos 问题
- 确保 principal 和 keytab 配置正确。
- 需在 BE 节点启动定时任务(如 crontab),每隔一定时间(如 12 小时),执行一次
kinit -kt your_principal your_keytab
命令。
Unknown type value: UNSUPPORTED
这是 Doris 2.0.2 版本和 Paimon 0.5 版本的一个兼容性问题,需要升级到 2.0.3 或更高版本解决,或自行 patch
访问对象存储(OSS、S3 等)报错文件系统不支持
在 2.0.5(含)之前的版本,用户需手动下载以下 jar 包并放置在
${DORIS_HOME}/be/lib/java_extensions/preload-extensions
目录下,重启 BE。- 访问 OSS:paimon-oss-0.6.0-incubating.jar
- 访问其他对象存储:paimon-s3-0.6.0-incubating.jar
2.0.6 之后的版本不再需要用户手动放置。