[jira] [Commented] (FLINK-10469) FileChannel may not write the whole buffer in a single call to FileChannel.write(Buffer buffer)
[ https://issues.apache.org/jira/browse/FLINK-10469?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633622#comment-16633622 ] Guowei Ma commented on FLINK-10469: --- +1 > FileChannel may not write the whole buffer in a single call to > FileChannel.write(Buffer buffer) > --- > > Key: FLINK-10469 > URL: https://issues.apache.org/jira/browse/FLINK-10469 > Project: Flink > Issue Type: Bug > Components: Core >Affects Versions: 1.4.1, 1.4.2, 1.5.3, 1.6.0, 1.6.1, 1.7.0, 1.5.4, 1.6.2 >Reporter: Yun Gao >Priority: Major > > Currently all the calls to _FileChannel.write(ByteBuffer src)_ assumes that > this method will not return before the whole buffer is written, like the one > in _AsynchronousFileIOChannel.write()._ > > However, this assumption may not be right for all the environments. We have > encountered the case that only part of a buffer was written on a cluster with > a high IO load, and the target file got messy. > > To fix this issue, I think we should add a utility method in the > org.apache.flink.util.IOUtils to ensure the whole buffer is written with a > loop,and replace all the calls to _FileChannel.write(ByteBuffer)_ with this > new method. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (FLINK-10469) FileChannel may not write the whole buffer in a single call to FileChannel.write(Buffer buffer)
Yun Gao created FLINK-10469: --- Summary: FileChannel may not write the whole buffer in a single call to FileChannel.write(Buffer buffer) Key: FLINK-10469 URL: https://issues.apache.org/jira/browse/FLINK-10469 Project: Flink Issue Type: Bug Components: Core Affects Versions: 1.5.4, 1.6.1, 1.6.0, 1.5.3, 1.4.2, 1.4.1, 1.7.0, 1.6.2 Reporter: Yun Gao Currently all the calls to _FileChannel.write(ByteBuffer src)_ assumes that this method will not return before the whole buffer is written, like the one in _AsynchronousFileIOChannel.write()._ However, this assumption may not be right for all the environments. We have encountered the case that only part of a buffer was written on a cluster with a high IO load, and the target file got messy. To fix this issue, I think we should add a utility method in the org.apache.flink.util.IOUtils to ensure the whole buffer is written with a loop,and replace all the calls to _FileChannel.write(ByteBuffer)_ with this new method. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (FLINK-9084) Upgrade cassandra-driver-core to 3.6.0 to introduce BusyPoolException retries to Cassandra Connector
[ https://issues.apache.org/jira/browse/FLINK-9084?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jacob Park updated FLINK-9084: -- Summary: Upgrade cassandra-driver-core to 3.6.0 to introduce BusyPoolException retries to Cassandra Connector (was: Upgrade cassandra-driver-core to 3.4.0 to introduce BusyPoolException retries to Cassandra Connector) > Upgrade cassandra-driver-core to 3.6.0 to introduce BusyPoolException retries > to Cassandra Connector > > > Key: FLINK-9084 > URL: https://issues.apache.org/jira/browse/FLINK-9084 > Project: Flink > Issue Type: Improvement > Components: Cassandra Connector >Reporter: Jacob Park >Assignee: Jacob Park >Priority: Minor > > As per [https://datastax-oss.atlassian.net/browse/SPARKC-503,] that exception > will be thrown by executeAsync() since at least 3.1.4 as executeAsync() is > fully async, so it throws that exception if the driver-side pool is busy. > This improvement has greatly improved the peek throughput of our Cassandra > Sink Connector. It also allowed us to get all the benefits of the latest > driver version. > I would like to contribute this feature back upstream. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Comment Edited] (FLINK-6966) Add maxParallelism and UIDs to all operators generated by the Table API / SQL
[ https://issues.apache.org/jira/browse/FLINK-6966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633442#comment-16633442 ] Fan weiwen edited comment on FLINK-6966 at 9/30/18 5:04 PM: this is my test sql {code:java} //代码占位符 i use redis sink but can test in kafka or other sink Asql select reqIp as factorContenta, count(*) as eCount, 60 * 60 as expire from kafkasource where uri in not null group by hop( rowtime, interval '2' second, interval '60' minute ), reqIp having count(*) >= 1 Bsql select uid as factorContentb, count(*) as eCount, 60 * 60 as expire from kafkasource where uri is not null group by hop( rowtime, interval '2' second, interval '60' minute ), uid having count(*) >= 1 {code} this is my test data {code:java} //代码占位符 { "code" : "200", "reqIp" : "656.19.173.34", "t" : 1537950912546, "uid" : 6630519, "uri" : "/test" } {code} first the job only Asql , start A sql redis has a key 656.19.173.34 then stop Asql and savepoint del redis key 656.19.173.34 now start Bsql from savepoint you can find it key 656.19.173.34 is create redis has Asql create key and Bsql create key this is wrong, Bsql fetch from Asql state was (Author: fanweiwen): this is my test sql {code:java} //代码占位符 i use redis sink but can test in kafka or other sink Asql select reqIp as factorContenta, count(*) as eCount, 60 * 60 as expire from kafkasource where uri in not null group by hop( rowtime, interval '2' second, interval '60' minute ), reqIp having count(*) >= 1 Bsql select uid as factorContentb, count(*) as eCount, 60 * 60 as expire from kafkasource where uri is not null group by hop( rowtime, interval '2' second, interval '60' minute ), uid having count(*) >= 1 {code} this is my test data {code:java} //代码占位符 { "code" : "200", "reqIp" : "656.19.173.34", "t" : 1537950912546, "uid" : 6630519, "uri" : "/test" } {code} first the job only Asql , start A sql redis has a key 656.19.173.34 then stop Asql and savepoint del redis key 656.19.173.34 now start Bsql from savepoint you can find it key 656.19.173.34 is create redis has Asql create key and Bsql create key this is wrong > Add maxParallelism and UIDs to all operators generated by the Table API / SQL > - > > Key: FLINK-6966 > URL: https://issues.apache.org/jira/browse/FLINK-6966 > Project: Flink > Issue Type: Improvement > Components: Table API SQL >Affects Versions: 1.4.0 >Reporter: Fabian Hueske >Priority: Major > > At the moment, the Table API does not assign UIDs and the max parallelism to > operators (except for operators with parallelism 1). > We should do that to avoid problems when rescaling or restarting jobs from > savepoints. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Comment Edited] (FLINK-6966) Add maxParallelism and UIDs to all operators generated by the Table API / SQL
[ https://issues.apache.org/jira/browse/FLINK-6966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633442#comment-16633442 ] Fan weiwen edited comment on FLINK-6966 at 9/30/18 5:02 PM: this is my test sql {code:java} //代码占位符 i use redis sink but can test in kafka or other sink Asql select reqIp as factorContenta, count(*) as eCount, 60 * 60 as expire from kafkasource where uri in not null group by hop( rowtime, interval '2' second, interval '60' minute ), reqIp having count(*) >= 1 Bsql select uid as factorContentb, count(*) as eCount, 60 * 60 as expire from kafkasource where uri is not null group by hop( rowtime, interval '2' second, interval '60' minute ), uid having count(*) >= 1 {code} this is my test data {code:java} //代码占位符 { "code" : "200", "reqIp" : "656.19.173.34", "t" : 1537950912546, "uid" : 6630519, "uri" : "/test" } {code} first the job only Asql , start A sql redis has a key 656.19.173.34 then stop Asql and savepoint del redis key 656.19.173.34 now start Bsql from savepoint you can find it key 656.19.173.34 is create redis has Asql create key and Bsql create key this is wrong was (Author: fanweiwen): this is my test sql {code:java} //代码占位符 i use redis sink but can test in kafka or other sink Asql select reqIp as factorContenta, count(*) as eCount, 60 * 60 as expire from xp_sso_biz_source where uri in not null group by hop( rowtime, interval '2' second, interval '60' minute ), reqIp having count(*) >= 1 Bsql select uid as factorContentb, count(*) as eCount, 60 * 60 as expire from xp_sso_biz_source where uri is not null group by hop( rowtime, interval '2' second, interval '60' minute ), uid having count(*) >= 1 {code} this is my test data {code:java} //代码占位符 { "code" : "200", "reqIp" : "656.19.173.34", "t" : 1537950912546, "uid" : 6630519, "uri" : "/test" } {code} first the job only Asql , start A sql redis has a key 656.19.173.34 then stop Asql and savepoint del redis key 656.19.173.34 now start Bsql from savepoint you can find it key 656.19.173.34 is create redis has Asql create key and Bsql create key this is wrong > Add maxParallelism and UIDs to all operators generated by the Table API / SQL > - > > Key: FLINK-6966 > URL: https://issues.apache.org/jira/browse/FLINK-6966 > Project: Flink > Issue Type: Improvement > Components: Table API SQL >Affects Versions: 1.4.0 >Reporter: Fabian Hueske >Priority: Major > > At the moment, the Table API does not assign UIDs and the max parallelism to > operators (except for operators with parallelism 1). > We should do that to avoid problems when rescaling or restarting jobs from > savepoints. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (FLINK-6966) Add maxParallelism and UIDs to all operators generated by the Table API / SQL
[ https://issues.apache.org/jira/browse/FLINK-6966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633442#comment-16633442 ] Fan weiwen commented on FLINK-6966: --- this is my test sql {code:java} //代码占位符 i use redis sink but can test in kafka or other sink Asql select reqIp as factorContenta, count(*) as eCount, 60 * 60 as expire from xp_sso_biz_source where uri in not null group by hop( rowtime, interval '2' second, interval '60' minute ), reqIp having count(*) >= 1 Bsql select uid as factorContentb, count(*) as eCount, 60 * 60 as expire from xp_sso_biz_source where uri is not null group by hop( rowtime, interval '2' second, interval '60' minute ), uid having count(*) >= 1 {code} this is my test data {code:java} //代码占位符 { "code" : "200", "reqIp" : "656.19.173.34", "t" : 1537950912546, "uid" : 6630519, "uri" : "/test" } {code} first the job only Asql , start A sql redis has a key 656.19.173.34 then stop Asql and savepoint del redis key 656.19.173.34 now start Bsql from savepoint you can find it key 656.19.173.34 is create redis has Asql create key and Bsql create key this is wrong > Add maxParallelism and UIDs to all operators generated by the Table API / SQL > - > > Key: FLINK-6966 > URL: https://issues.apache.org/jira/browse/FLINK-6966 > Project: Flink > Issue Type: Improvement > Components: Table API SQL >Affects Versions: 1.4.0 >Reporter: Fabian Hueske >Priority: Major > > At the moment, the Table API does not assign UIDs and the max parallelism to > operators (except for operators with parallelism 1). > We should do that to avoid problems when rescaling or restarting jobs from > savepoints. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (FLINK-6966) Add maxParallelism and UIDs to all operators generated by the Table API / SQL
[ https://issues.apache.org/jira/browse/FLINK-6966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633423#comment-16633423 ] Fan weiwen commented on FLINK-6966: --- [~hequn8128] it not like this i think it independent of the field uids is not unique this is a very easy question to reproduce can try it > Add maxParallelism and UIDs to all operators generated by the Table API / SQL > - > > Key: FLINK-6966 > URL: https://issues.apache.org/jira/browse/FLINK-6966 > Project: Flink > Issue Type: Improvement > Components: Table API SQL >Affects Versions: 1.4.0 >Reporter: Fabian Hueske >Priority: Major > > At the moment, the Table API does not assign UIDs and the max parallelism to > operators (except for operators with parallelism 1). > We should do that to avoid problems when rescaling or restarting jobs from > savepoints. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (FLINK-10468) Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor
[ https://issues.apache.org/jira/browse/FLINK-10468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633416#comment-16633416 ] Ted Yu commented on FLINK-10468: That was why I started the issue title with Potential. Even if this is the case, assigning {{channels}} and breaking would make the code easier to understand for other people. Or, a comment should be added stating the fact. > Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor > -- > > Key: FLINK-10468 > URL: https://issues.apache.org/jira/browse/FLINK-10468 > Project: Flink > Issue Type: Bug >Reporter: Ted Yu >Priority: Minor > > Here is related code: > {code} > switch (strategy) { > case PARTITION_CUSTOM: > extractedKeys = new Object[1]; > case FORWARD: > {code} > It seems a 'break' is missing prior to FORWARD case. > {code} > if (strategy == ShipStrategyType.PARTITION_CUSTOM && partitioner == null) > { > throw new NullPointerException("Partitioner must not be null when the > ship strategy is set to custom partitioning."); > } > {code} > Since the above check is for PARTITION_CUSTOM, it seems we can place the > check in the switch statement. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (FLINK-10468) Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor
[ https://issues.apache.org/jira/browse/FLINK-10468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633413#comment-16633413 ] tison commented on FLINK-10468: --- I just have taken a look at that code and think it is intended since {{PARTITION_CUSTOM}} means that partitioning using a custom partitioner so it sets a {{extractedKeys}}, but all of the strategy need a {{channels}}. However, I am not quite familiar with this code and think [~zjwang], [~piwaniuk] and [~NicoK] would know more about it. And if they are stand the same way as me, we can close this issue as won't fix. > Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor > -- > > Key: FLINK-10468 > URL: https://issues.apache.org/jira/browse/FLINK-10468 > Project: Flink > Issue Type: Bug >Reporter: Ted Yu >Priority: Minor > > Here is related code: > {code} > switch (strategy) { > case PARTITION_CUSTOM: > extractedKeys = new Object[1]; > case FORWARD: > {code} > It seems a 'break' is missing prior to FORWARD case. > {code} > if (strategy == ShipStrategyType.PARTITION_CUSTOM && partitioner == null) > { > throw new NullPointerException("Partitioner must not be null when the > ship strategy is set to custom partitioning."); > } > {code} > Since the above check is for PARTITION_CUSTOM, it seems we can place the > check in the switch statement. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Assigned] (FLINK-10468) Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor
[ https://issues.apache.org/jira/browse/FLINK-10468?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] tison reassigned FLINK-10468: - Assignee: (was: tison) > Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor > -- > > Key: FLINK-10468 > URL: https://issues.apache.org/jira/browse/FLINK-10468 > Project: Flink > Issue Type: Bug >Reporter: Ted Yu >Priority: Minor > > Here is related code: > {code} > switch (strategy) { > case PARTITION_CUSTOM: > extractedKeys = new Object[1]; > case FORWARD: > {code} > It seems a 'break' is missing prior to FORWARD case. > {code} > if (strategy == ShipStrategyType.PARTITION_CUSTOM && partitioner == null) > { > throw new NullPointerException("Partitioner must not be null when the > ship strategy is set to custom partitioning."); > } > {code} > Since the above check is for PARTITION_CUSTOM, it seems we can place the > check in the switch statement. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (FLINK-6966) Add maxParallelism and UIDs to all operators generated by the Table API / SQL
[ https://issues.apache.org/jira/browse/FLINK-6966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633410#comment-16633410 ] Hequn Cheng commented on FLINK-6966: Hi [~fanweiwen] , this task has not been started. In your case, it seems that jobB shares same uids with jobA. So, jobB can restore from the state if field b and field c share same filed type. > Add maxParallelism and UIDs to all operators generated by the Table API / SQL > - > > Key: FLINK-6966 > URL: https://issues.apache.org/jira/browse/FLINK-6966 > Project: Flink > Issue Type: Improvement > Components: Table API SQL >Affects Versions: 1.4.0 >Reporter: Fabian Hueske >Priority: Major > > At the moment, the Table API does not assign UIDs and the max parallelism to > operators (except for operators with parallelism 1). > We should do that to avoid problems when rescaling or restarting jobs from > savepoints. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Assigned] (FLINK-10468) Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor
[ https://issues.apache.org/jira/browse/FLINK-10468?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] tison reassigned FLINK-10468: - Assignee: tison > Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor > -- > > Key: FLINK-10468 > URL: https://issues.apache.org/jira/browse/FLINK-10468 > Project: Flink > Issue Type: Bug >Reporter: Ted Yu >Assignee: tison >Priority: Minor > > Here is related code: > {code} > switch (strategy) { > case PARTITION_CUSTOM: > extractedKeys = new Object[1]; > case FORWARD: > {code} > It seems a 'break' is missing prior to FORWARD case. > {code} > if (strategy == ShipStrategyType.PARTITION_CUSTOM && partitioner == null) > { > throw new NullPointerException("Partitioner must not be null when the > ship strategy is set to custom partitioning."); > } > {code} > Since the above check is for PARTITION_CUSTOM, it seems we can place the > check in the switch statement. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (FLINK-10468) Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor
Ted Yu created FLINK-10468: -- Summary: Potential missing break for PARTITION_CUSTOM in OutputEmitter ctor Key: FLINK-10468 URL: https://issues.apache.org/jira/browse/FLINK-10468 Project: Flink Issue Type: Bug Reporter: Ted Yu Here is related code: {code} switch (strategy) { case PARTITION_CUSTOM: extractedKeys = new Object[1]; case FORWARD: {code} It seems a 'break' is missing prior to FORWARD case. {code} if (strategy == ShipStrategyType.PARTITION_CUSTOM && partitioner == null) { throw new NullPointerException("Partitioner must not be null when the ship strategy is set to custom partitioning."); } {code} Since the above check is for PARTITION_CUSTOM, it seems we can place the check in the switch statement. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (FLINK-9932) If task executor offer slot to job master timeout the first time, the slot will leak
[ https://issues.apache.org/jira/browse/FLINK-9932?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633307#comment-16633307 ] ASF GitHub Bot commented on FLINK-9932: --- shuai-xu opened a new pull request #6780: [FLINK-9932] [runtime] fix slot leak when task executor offer slot to job master timeout URL: https://github.com/apache/flink/pull/6780 ## What is the purpose of the change *(For example: This pull request fix that the slots in task executor will leak if task executor fail to offer it to job master due to rpc timeout.)* ## Verifying this change - *Add a test in TaskExecutorTest* ## Does this pull request potentially affect one of the following parts: - Dependencies (does it add or upgrade a dependency): (no) - The public API, i.e., is any changed class annotated with `@Public(Evolving)`: (no) - The serializers: (no) - The runtime per-record code paths (performance sensitive): (no ) - Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Yarn/Mesos, ZooKeeper: (no) - The S3 file system connector: (no) ## Documentation - Does this pull request introduce a new feature? (no) - If yes, how is the feature documented? (not applicable) This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org > If task executor offer slot to job master timeout the first time, the slot > will leak > > > Key: FLINK-9932 > URL: https://issues.apache.org/jira/browse/FLINK-9932 > Project: Flink > Issue Type: Bug > Components: Cluster Management >Affects Versions: 1.5.0 >Reporter: shuai.xu >Assignee: shuai.xu >Priority: Major > Labels: pull-request-available > > When task executor offer slot to job master, it will first mark the slot as > active. > If the offer slot call timeout, the task executor will try to call > offerSlotsToJobManager again, > but it will only offer the slot in ALLOCATED state. As the slot has already > be mark ACTIVE, it will never be offered and this will cause slot leak. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (FLINK-9932) If task executor offer slot to job master timeout the first time, the slot will leak
[ https://issues.apache.org/jira/browse/FLINK-9932?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] ASF GitHub Bot updated FLINK-9932: -- Labels: pull-request-available (was: ) > If task executor offer slot to job master timeout the first time, the slot > will leak > > > Key: FLINK-9932 > URL: https://issues.apache.org/jira/browse/FLINK-9932 > Project: Flink > Issue Type: Bug > Components: Cluster Management >Affects Versions: 1.5.0 >Reporter: shuai.xu >Assignee: shuai.xu >Priority: Major > Labels: pull-request-available > > When task executor offer slot to job master, it will first mark the slot as > active. > If the offer slot call timeout, the task executor will try to call > offerSlotsToJobManager again, > but it will only offer the slot in ALLOCATED state. As the slot has already > be mark ACTIVE, it will never be offered and this will cause slot leak. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[GitHub] shuai-xu opened a new pull request #6780: [FLINK-9932] [runtime] fix slot leak when task executor offer slot to job master timeout
shuai-xu opened a new pull request #6780: [FLINK-9932] [runtime] fix slot leak when task executor offer slot to job master timeout URL: https://github.com/apache/flink/pull/6780 ## What is the purpose of the change *(For example: This pull request fix that the slots in task executor will leak if task executor fail to offer it to job master due to rpc timeout.)* ## Verifying this change - *Add a test in TaskExecutorTest* ## Does this pull request potentially affect one of the following parts: - Dependencies (does it add or upgrade a dependency): (no) - The public API, i.e., is any changed class annotated with `@Public(Evolving)`: (no) - The serializers: (no) - The runtime per-record code paths (performance sensitive): (no ) - Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Yarn/Mesos, ZooKeeper: (no) - The S3 file system connector: (no) ## Documentation - Does this pull request introduce a new feature? (no) - If yes, how is the feature documented? (not applicable) This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[jira] [Commented] (FLINK-6966) Add maxParallelism and UIDs to all operators generated by the Table API / SQL
[ https://issues.apache.org/jira/browse/FLINK-6966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16633255#comment-16633255 ] Fan weiwen commented on FLINK-6966: --- [~sunjincheng121] This is the problem, is it solved? i have a similar problem the job has two process one table one sink for example A // view select a,b, count(*) as acnt from kafka group by hop( rowtime, interval '2' second, interval '60' minute ), a, b Table table = tableEnv.sqlQuery(sql); tableEnv.registerTable(Aname, table); // sink sinksql = select a,b,ant from Aname Table table = tableEnv.sqlQuery(sinksql); DataStream ds = tableEnv.toAppendStream(table, Row.class); ds.addSink(...); B // view sql = select a,c, count(*) as bnt from kafka group by hop( rowtime, interval '2' second, interval '60' minute ), a, c Table table = tableEnv.sqlQuery(sql); tableEnv.registerTable(Bname, table); // sink sinksql = select a,c,bnt from Bname Table table = tableEnv.sqlQuery(sinksql); DataStream ds = tableEnv.toAppendStream(table, Row.class); ds.addSink(...); - now the job start A, stop B then stop A , cancel and savepoint start B , from savepoint B sink sinksql = select a,c,bnt from Bname problem is B sink fetch data from A state State is not maintained only, not uid and operator id english is pool , can chinese exchange ? > Add maxParallelism and UIDs to all operators generated by the Table API / SQL > - > > Key: FLINK-6966 > URL: https://issues.apache.org/jira/browse/FLINK-6966 > Project: Flink > Issue Type: Improvement > Components: Table API SQL >Affects Versions: 1.4.0 >Reporter: Fabian Hueske >Priority: Major > > At the moment, the Table API does not assign UIDs and the max parallelism to > operators (except for operators with parallelism 1). > We should do that to avoid problems when rescaling or restarting jobs from > savepoints. -- This message was sent by Atlassian JIRA (v7.6.3#76005)