Hi Jing,
Yes, FIRST is a UDAF.
I've been trying to reproduce this locally without success so far.
The query itself has more fields and aggregates. Once I can reproduce this
locally I'll try to narrow down the problematic field and share more
information.
On Tue, Jul 27, 2021, 05:17 JING ZHANG
Hi Mason,
In rocksDB, one state is corresponding to a column family and we could
aggregate all RocksDB native metrics per column family. If my understanding is
right, are you hoping that all state latency metrics for a particular state
could be aggregated per state level?
Best
Yun Tang
是的,类似的
On Tue, Jul 27, 2021 at 10:42 AM lixin58...@163.com
wrote:
> 你好,
>
> 感谢回复,看了下这个文档,提到对于parquet这种列式文件只会使用onCheckpointRollPolicy,也就是只在做检查点时会滚动。flink
> filesystem table这块的parquet列式文件写入是不是也这样呢?
>
>
你好,
感谢回复,看了下这个文档,提到对于parquet这种列式文件只会使用onCheckpointRollPolicy,也就是只在做检查点时会滚动。flink
filesystem table这块的parquet列式文件写入是不是也这样呢?
https://ci.apache.org/projects/flink/flink-docs-release-1.12/zh/dev/connectors/streamfile_sink.html
lixin58...@163.com
发件人: Jingsong Li
发送时间: 2021-07-27 10:30
parquet因为它会在内存中攒buffer,所以文件的file-size并不能很精确。。只能等它flush了才会生效。
On Sun, Jul 25, 2021 at 9:47 AM lixin58...@163.com
wrote:
> 大家好,
>
> 检查点配的是120s,滚动时长800s,滚动大小1kb,并行度配的2
>
>
> 不过在跑的过程中发现不管写入的多快,同时只存在一个in-progress文件,且最终生成的文件是严格按照120s生成的,这个很奇怪,似乎只有按检查点滚动生效了,与json格式的不一样。真的是这样吗?不过看官方文档没有这样说
>
> 求大佬们解惑!
Hi,caizhi,非常感谢你的回复!
在KafkaDynamicTableFactory.java的createDynamicTableSink(Context
context)方法开始,我通过context.getObjectIdentifier().getObjectName()获取到sinkTableName。因为ObjectIdentifier类就唯一标识了一个表,它包括catalogName、databaseName和objectName,分别代表catalog名、数据库名和表名。之后,我通过构造传入到了FlinkKafkaProducer,然后就可以使用了。
Hi, Ivan
My gut feeling is that it is related to FLINK-22535. Could @Yun Gao
take another look? If that is the case, you can upgrade to 1.13.1.
Best,
Yangze Guo
On Tue, Jul 27, 2021 at 9:41 AM Ivan Yang wrote:
>
> Dear Flink experts,
>
> We recently ran into an issue during a job cancellation
Hi Yuval,
I run a similar SQL (without `FIRST` aggregate function), there is nothing
wrong.
`FIRST` is a custom aggregate function? Would you please check if there is
a drawback in `FIRST`? Whether the query could run without `FIRST`?
Best,
JING ZHANG
Yuval Itzchakov 于2021年7月27日周二 上午12:29写道:
>
Dear Flink experts,
We recently ran into an issue during a job cancellation after upgraded to 1.13.
After we issue a cancel (from Flink console or flink cancel {jobid}), a few
subtasks stuck in cancelling state. Once it gets to that situation, the
behavior is consistent. Those “cancelling
We have been using the state backend latency tracking metrics from Flink
1.13. To make metrics aggregation easier, could there be a config to expose
something like `state.backend.rocksdb.metrics.column-family-as-variable`
that rocksdb provides to do aggregation across column families.
In this
Hi,
*Setup:*
1 JM,
1 TM,
Flink 1.13.1
RocksDBStateBackend.
I have a query with the rough sketch of the following:
SELECT CAST(TUMBLE_START(event_time, INTERVAL '2' MINUTE) AS TIMESTAMP)
START_TIME
CAST(TUMBLE_END(event_time, INTERVAL '2' MINUTE) AS
TIMESTAMP) END_TIME
Hi,
It is recommended to package your application with all the
dependencies into a single file [1].
And according to the kafka-connector documentation [2]:
if you are using Kafka source, flink-connector-base is also required
as dependency:
org.apache.flink
flink-connector-base
Hello,
Could you check that TMs didn't fail and therefore unregistered KV
states and are still running at the time of the query?
Probably after changing the memory settings there is another error
that is reported later than the state is unregistered.
Regards,
Roman
On Sat, Jul 24, 2021 at 12:50
1.13不再使用is_generic来标记是不是hive表了 (改成了'connector'='hive')
,文档需要更新一下。不过还是建议DDL操作hive元数据时用hive dialect。
On Mon, Jul 26, 2021 at 5:00 PM Asahi Lee <978466...@qq.com.invalid> wrote:
> 我使用flink 1.12.0版本功能是好的
>
>
>
>
> --原始邮件--
> 发件人:
>
确切的说我已经遇到了这种问题,就是读入了penging或者in-progress状态的文件,然后就报FileNotFoundException了,
是偶现的。那这种情况下除了重新编译FileInputFormat.java文件,过滤掉这两种状态的文件外,flink中还有没有提供什么其他解决办法?
在 2021-07-26 16:34:19,"Michael Ran" 写道:
你尝试下 用java 读取文件过程中,改文件名 或者移动文件。会异常
在 2021-07-26 15:19:39,"wanggaoliang" <18838915...@163.com> 写道:
??flink 1.12.0??
----
??:
"user-zh"
你好,
是否能尝试一下用hive dialect建表呢?
On Mon, Jul 26, 2021 at 2:44 PM Asahi Lee <978466...@qq.com.invalid> wrote:
> hi!
> 我使用flink 1.13.1版本,通过sql创建hive表,程序正常运行,表创建成功,但是没有列信息;我的程序如下:
> 我连接的是hive 2.3.6版本,使用flink-sql-connector-hive-2.3.6依赖包。
>
>
> package com.meritdata.cloud.flink.test;
>
>
> import
额,混用多个UDF没法,本身就依赖冲突了,一般公共的UDF我们都统一的。 私人的和 公共的冲突了- -再单独和他们讨论
在 2021-07-26 11:43:49,"silence" 写道:
>就是单独引用的啊,但任务逻辑比较复杂时会同时混用多个udf这个是没法避免的啊
>
>
>--
>发件人:Michael Ran
>发送时间:2021年7月23日(星期五) 17:42
>收件人:user-zh ; silence
>主 题:Re:回复:flink
你尝试下 用java 读取文件过程中,改文件名 或者移动文件。会异常
在 2021-07-26 15:19:39,"wanggaoliang" <18838915...@163.com> 写道:
在 2021-07-26 14:56:00,"wanggaoliang" <18838915...@163.com> 写道:
flink读取hdfs文件时,FileInputFormat.java类中,acceptFile()方法默认过滤掉了以"_"和"."开头的文件,那如果读入了in-progress文件和.pending文件,
Hi,all
使用zeppelin提交sqljob的时候发现在 flink jobmanager 中首先会打印以下日志,
2021-07-26 15:54:36,929 [Thread-6575] INFO
org.apache.hadoop.hdfs.protocol.datatransfer.sasl.SaslDataTransferClient
Hello,
If you are consuming from a single stream, you can use the shard ID to
achieve a better distribution. Since the shard IDs are assigned
incrementally like so:
- shardId-
- shardId-0001
- shardId-0002
- etc
You can substring the prefix and
flink读取hdfs文件时,FileInputFormat.java类中,acceptFile()方法默认过滤掉了以"_"和"."开头的文件,那如果读入了in-progress文件和.pending文件,
在执行open()方法时,突然in-progress或.pending文件状态发生了改变而导致原来的文件路径消失,会不会出现什么问题?或者是有没有可能出现这种情况?
hi!
??flink
1.13.1??sqlhive
??hive 2.3.6??flink-sql-connector-hive-2.3.6
package com.meritdata.cloud.flink.test;
import org.apache.flink.table.api.EnvironmentSettings;
import
23 matches
Mail list logo