[jira] [Created] (HUDI-8074) Improve comaction operator shuffle rebanlance
Danny Chen created HUDI-8074: Summary: Improve comaction operator shuffle rebanlance Key: HUDI-8074 URL: https://issues.apache.org/jira/browse/HUDI-8074 Project: Apache Hudi Issue Type: Improvement Reporter: Danny Chen -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Closed] (HUDI-8074) Improve comaction operator shuffle rebanlance
[ https://issues.apache.org/jira/browse/HUDI-8074?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-8074. Resolution: Fixed Fixed via master branch: 35c00daaf871a6c1b87d6a440832d60f9b26ee14 > Improve comaction operator shuffle rebanlance > -- > > Key: HUDI-8074 > URL: https://issues.apache.org/jira/browse/HUDI-8074 > Project: Apache Hudi > Issue Type: Improvement >Reporter: Danny Chen >Priority: Major > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Created] (HUDI-8077) Fix the incremental cleaning to base on completion time
Danny Chen created HUDI-8077: Summary: Fix the incremental cleaning to base on completion time Key: HUDI-8077 URL: https://issues.apache.org/jira/browse/HUDI-8077 Project: Apache Hudi Issue Type: Sub-task Reporter: Danny Chen Fix For: 1.0.0 -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Updated] (HUDI-8077) Fix the incremental cleaning to base on completion time
[ https://issues.apache.org/jira/browse/HUDI-8077?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-8077: - Description: Currently, the incremental cleaning will remember a marker instant of last retained in the commit metadata, both the marker and the filtering instant on fs view are start times(instant times), while this is okay for most of the cases because we actually have some buffer time for cleaning(30 commits retained by default), but if the user sets up a very redical strategy like clean for every commits, then there might be isses in NB-CC mode: A instant that starts very early but finished recently might be skipped by the cleaning table service. > Fix the incremental cleaning to base on completion time > --- > > Key: HUDI-8077 > URL: https://issues.apache.org/jira/browse/HUDI-8077 > Project: Apache Hudi > Issue Type: Sub-task >Reporter: Danny Chen >Priority: Major > Fix For: 1.0.0 > > > Currently, the incremental cleaning will remember a marker instant of last > retained in the commit metadata, both the marker and the filtering instant on > fs view are start times(instant times), while this is okay for most of the > cases because we actually have some buffer time for cleaning(30 commits > retained by default), but if the user sets up a very redical strategy like > clean for every commits, then there might be isses in NB-CC mode: > A instant that starts very early but finished recently might be skipped by > the cleaning table service. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Closed] (HUDI-7873) Remove getStorage method from HoodieReaderContext
[ https://issues.apache.org/jira/browse/HUDI-7873?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-7873. Fix Version/s: 1.0.0 Resolution: Fixed Fixed via master branch: 64f546b8f0cae70793a6150170a649bad8e0e146 > Remove getStorage method from HoodieReaderContext > - > > Key: HUDI-7873 > URL: https://issues.apache.org/jira/browse/HUDI-7873 > Project: Apache Hudi > Issue Type: Improvement >Reporter: Jonathan Vexler >Assignee: Jonathan Vexler >Priority: Major > Labels: pull-request-available > Fix For: 1.0.0 > > > All implementations of the method were the same, and it was only used by a > test method becuase storage is passed as a param to the fg reader. > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Updated] (HUDI-8097) Schema evolution setting from hudi-defaults.conf is ignored while altering column in Spark
[ https://issues.apache.org/jira/browse/HUDI-8097?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-8097: - Fix Version/s: 1.0.0 > Schema evolution setting from hudi-defaults.conf is ignored while altering > column in Spark > -- > > Key: HUDI-8097 > URL: https://issues.apache.org/jira/browse/HUDI-8097 > Project: Apache Hudi > Issue Type: Bug >Reporter: Vova Kolmakov >Assignee: Vova Kolmakov >Priority: Minor > Labels: pull-request-available > Fix For: 1.0.0 > > > Value of `hoodie.schema.on.read.enable` from externalized config > `hudi-defaults.conf` is not taken into consideration while processing > commands such as `alter table change column .. ` in Spark. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Updated] (HUDI-8097) Schema evolution setting from hudi-defaults.conf is ignored while altering column in Spark
[ https://issues.apache.org/jira/browse/HUDI-8097?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-8097: - Status: Open (was: In Progress) > Schema evolution setting from hudi-defaults.conf is ignored while altering > column in Spark > -- > > Key: HUDI-8097 > URL: https://issues.apache.org/jira/browse/HUDI-8097 > Project: Apache Hudi > Issue Type: Bug >Reporter: Vova Kolmakov >Assignee: Vova Kolmakov >Priority: Minor > Labels: pull-request-available > Fix For: 1.0.0 > > > Value of `hoodie.schema.on.read.enable` from externalized config > `hudi-defaults.conf` is not taken into consideration while processing > commands such as `alter table change column .. ` in Spark. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Closed] (HUDI-8097) Schema evolution setting from hudi-defaults.conf is ignored while altering column in Spark
[ https://issues.apache.org/jira/browse/HUDI-8097?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-8097. Resolution: Fixed Fixed via master branch: e4e5bf0232093bfba25f1769fb63258e9f816a8f > Schema evolution setting from hudi-defaults.conf is ignored while altering > column in Spark > -- > > Key: HUDI-8097 > URL: https://issues.apache.org/jira/browse/HUDI-8097 > Project: Apache Hudi > Issue Type: Bug >Reporter: Vova Kolmakov >Assignee: Vova Kolmakov >Priority: Minor > Labels: pull-request-available > Fix For: 1.0.0 > > > Value of `hoodie.schema.on.read.enable` from externalized config > `hudi-defaults.conf` is not taken into consideration while processing > commands such as `alter table change column .. ` in Spark. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Updated] (HUDI-8104) Hive on spark select mor rt table, org.apache.hudi.hadoop.hive.HoodieCombineRealtimeFileSplit cannot be cast to org.apache.hadoop.hive.shims.HadoopShimsSecure
[ https://issues.apache.org/jira/browse/HUDI-8104?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-8104: - Fix Version/s: 1.0.0 > Hive on spark select mor rt table, > org.apache.hudi.hadoop.hive.HoodieCombineRealtimeFileSplit cannot be cast to > org.apache.hadoop.hive.shims.HadoopShimsSecure > -- > > Key: HUDI-8104 > URL: https://issues.apache.org/jira/browse/HUDI-8104 > Project: Apache Hudi > Issue Type: Bug >Reporter: felixzh >Priority: Critical > Labels: pull-request-available > Fix For: 1.0.0 > > Original Estimate: 1h > Remaining Estimate: 1h > > Hive on spark select mor rt table, ERROR: > org.apache.hudi.hadoop.hive.HoodieCombineRealtimeFileSplit cannot be cast to > org.apache.hadoop.hive.shims.HadoopShimsSecure > The reason for the error is > HoodieParquetRealtimeInputFormat's job not set correctly hudi.hive.realtime -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Closed] (HUDI-8104) Hive on spark select mor rt table, org.apache.hudi.hadoop.hive.HoodieCombineRealtimeFileSplit cannot be cast to org.apache.hadoop.hive.shims.HadoopShimsSecure
[ https://issues.apache.org/jira/browse/HUDI-8104?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-8104. Resolution: Fixed Fixed via master branch: b6bad403fa97479c44112b741007a08102617ab3 > Hive on spark select mor rt table, > org.apache.hudi.hadoop.hive.HoodieCombineRealtimeFileSplit cannot be cast to > org.apache.hadoop.hive.shims.HadoopShimsSecure > -- > > Key: HUDI-8104 > URL: https://issues.apache.org/jira/browse/HUDI-8104 > Project: Apache Hudi > Issue Type: Bug >Reporter: felixzh >Priority: Critical > Labels: pull-request-available > Fix For: 1.0.0 > > Original Estimate: 1h > Remaining Estimate: 1h > > Hive on spark select mor rt table, ERROR: > org.apache.hudi.hadoop.hive.HoodieCombineRealtimeFileSplit cannot be cast to > org.apache.hadoop.hive.shims.HadoopShimsSecure > The reason for the error is > HoodieParquetRealtimeInputFormat's job not set correctly hudi.hive.realtime -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Closed] (HUDI-8112) Fix TestHoodieActiveTimeline#testTimelineGetOperations test logic error
[ https://issues.apache.org/jira/browse/HUDI-8112?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-8112. Resolution: Fixed Fixed via master branch: e1f70fda5890cff44b3e44eb362c5a411f4174f3 > Fix TestHoodieActiveTimeline#testTimelineGetOperations test logic error > --- > > Key: HUDI-8112 > URL: https://issues.apache.org/jira/browse/HUDI-8112 > Project: Apache Hudi > Issue Type: Improvement >Reporter: bradley >Priority: Major > Labels: pull-request-available > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Updated] (HUDI-8112) Fix TestHoodieActiveTimeline#testTimelineGetOperations test logic error
[ https://issues.apache.org/jira/browse/HUDI-8112?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-8112: - Fix Version/s: 1.0.0 > Fix TestHoodieActiveTimeline#testTimelineGetOperations test logic error > --- > > Key: HUDI-8112 > URL: https://issues.apache.org/jira/browse/HUDI-8112 > Project: Apache Hudi > Issue Type: Improvement >Reporter: bradley >Priority: Major > Labels: pull-request-available > Fix For: 1.0.0 > > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Closed] (HUDI-8070) Support Flink 1.19 in Hudi
[ https://issues.apache.org/jira/browse/HUDI-8070?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-8070. Resolution: Fixed Fixed via master branch: d550b46d72488f92abfb234f919506e76900d69f > Support Flink 1.19 in Hudi > -- > > Key: HUDI-8070 > URL: https://issues.apache.org/jira/browse/HUDI-8070 > Project: Apache Hudi > Issue Type: Task > Components: flink >Reporter: Zhenqiu Huang >Assignee: Zhenqiu Huang >Priority: Major > Labels: pull-request-available > Fix For: 1.0.0 > > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Closed] (HUDI-8078) Persisting writestatus to optimize the writes
[ https://issues.apache.org/jira/browse/HUDI-8078?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-8078. Resolution: Fixed Fixed via master branch: e0ef86421993c1664da5cbb3ab7de7e87f16cb49 > Persisting writestatus to optimize the writes > -- > > Key: HUDI-8078 > URL: https://issues.apache.org/jira/browse/HUDI-8078 > Project: Apache Hudi > Issue Type: Improvement > Components: writer-core >Reporter: Aditya Goenka >Priority: Critical > Labels: pull-request-available, spark > Fix For: 1.1.0, 0.15.1 > > > Details on this issue - [https://github.com/apache/hudi/issues/11741] -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Assigned] (HUDI-8078) Persisting writestatus to optimize the writes
[ https://issues.apache.org/jira/browse/HUDI-8078?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen reassigned HUDI-8078: Assignee: Danny Chen > Persisting writestatus to optimize the writes > -- > > Key: HUDI-8078 > URL: https://issues.apache.org/jira/browse/HUDI-8078 > Project: Apache Hudi > Issue Type: Improvement > Components: writer-core >Reporter: Aditya Goenka >Assignee: Danny Chen >Priority: Critical > Labels: pull-request-available, spark > Fix For: 1.1.0, 0.15.1 > > > Details on this issue - [https://github.com/apache/hudi/issues/11741] -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Updated] (HUDI-6788) Integrate FileGroupReader with MergeOnReadInputFormat for Flink
[ https://issues.apache.org/jira/browse/HUDI-6788?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-6788: - Description: The existing > Integrate FileGroupReader with MergeOnReadInputFormat for Flink > --- > > Key: HUDI-6788 > URL: https://issues.apache.org/jira/browse/HUDI-6788 > Project: Apache Hudi > Issue Type: New Feature >Reporter: Ethan Guo >Assignee: Zhenqiu Huang >Priority: Blocker > Fix For: 1.0.0 > > > The existing -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Updated] (HUDI-6788) Integrate FileGroupReader with MergeOnReadInputFormat for Flink
[ https://issues.apache.org/jira/browse/HUDI-6788?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-6788: - Description: The existing MergeOnReadInputFormat implements different iterators for all kinds of read more: incremental read, read optimized view, snapshot view etc. While for better performance and code evolving, we can integrate the new FileGroupReader, the main difference is that the FileGroupReader capsulate the file slice logs and parquet merging logic, so each iterator can ease the redundant work for quering the fs view and comprising the file slices. We can integrate step by step for different read views: 1. snapshot queries 2. read optimized queries 3. skip merge queries For usability and smoth evolving, we should add a flag for the new reader, the old code path should be kept there for 1 or 2 releases. The major work AIs includes: 1. implement the HoodieFlinkRecord akka to the HoodieSparkRecord; 2. implement the Flink specific FileGroupReader with the HoodieFlinkRecord; 3. Flink implements the snapshot queries using the file group reader; 4. Flink implements the read optimized queries using the file group reader; 5. Flink implements the skip merge queries using the file group reader. was:The existing > Integrate FileGroupReader with MergeOnReadInputFormat for Flink > --- > > Key: HUDI-6788 > URL: https://issues.apache.org/jira/browse/HUDI-6788 > Project: Apache Hudi > Issue Type: New Feature >Reporter: Ethan Guo >Assignee: Zhenqiu Huang >Priority: Blocker > Fix For: 1.0.0 > > > The existing MergeOnReadInputFormat implements different iterators for all > kinds of read more: incremental read, read optimized view, snapshot view etc. > While for better performance and code evolving, we can integrate the new > FileGroupReader, the main difference is that the FileGroupReader capsulate > the file slice logs and parquet merging logic, so each iterator can ease the > redundant work for quering the fs view and comprising the file slices. > We can integrate step by step for different read views: 1. snapshot queries > 2. read optimized queries 3. skip merge queries > For usability and smoth evolving, we should add a flag for the new reader, > the old code path should be kept there for 1 or 2 releases. > The major work AIs includes: > 1. implement the HoodieFlinkRecord akka to the HoodieSparkRecord; > 2. implement the Flink specific FileGroupReader with the HoodieFlinkRecord; > 3. Flink implements the snapshot queries using the file group reader; > 4. Flink implements the read optimized queries using the file group reader; > 5. Flink implements the skip merge queries using the file group reader. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Created] (HUDI-8118) Implement the Flink specific FileGroupReader with the HoodieFlinkRecord
Danny Chen created HUDI-8118: Summary: Implement the Flink specific FileGroupReader with the HoodieFlinkRecord Key: HUDI-8118 URL: https://issues.apache.org/jira/browse/HUDI-8118 Project: Apache Hudi Issue Type: Sub-task Components: flink Reporter: Danny Chen Assignee: Zhenqiu Huang Fix For: 1.0.0 -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Created] (HUDI-8117) Implement the HoodieFlinkRecord akka to the HoodieSparkRecord
Danny Chen created HUDI-8117: Summary: Implement the HoodieFlinkRecord akka to the HoodieSparkRecord Key: HUDI-8117 URL: https://issues.apache.org/jira/browse/HUDI-8117 Project: Apache Hudi Issue Type: Sub-task Components: flink Reporter: Danny Chen Assignee: Zhenqiu Huang Fix For: 1.0.0 -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Created] (HUDI-8119) Flink implements the snapshot queries using the file group reader
Danny Chen created HUDI-8119: Summary: Flink implements the snapshot queries using the file group reader Key: HUDI-8119 URL: https://issues.apache.org/jira/browse/HUDI-8119 Project: Apache Hudi Issue Type: Sub-task Components: flink Reporter: Danny Chen Assignee: Zhenqiu Huang Fix For: 1.0.0 -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Created] (HUDI-8120) Flink implements the read optimized queries using the file group reader
Danny Chen created HUDI-8120: Summary: Flink implements the read optimized queries using the file group reader Key: HUDI-8120 URL: https://issues.apache.org/jira/browse/HUDI-8120 Project: Apache Hudi Issue Type: Sub-task Components: flink Reporter: Danny Chen Assignee: Zhenqiu Huang Fix For: 1.0.0 -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Created] (HUDI-8121) Flink implements the skip merge queries using the file group reader
Danny Chen created HUDI-8121: Summary: Flink implements the skip merge queries using the file group reader Key: HUDI-8121 URL: https://issues.apache.org/jira/browse/HUDI-8121 Project: Apache Hudi Issue Type: Sub-task Components: flink Reporter: Danny Chen Assignee: Zhenqiu Huang Fix For: 1.0.0 -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Assigned] (HUDI-8136) New instant time generation for Flink streaming pipeline
[ https://issues.apache.org/jira/browse/HUDI-8136?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen reassigned HUDI-8136: Assignee: Danny Chen > New instant time generation for Flink streaming pipeline > > > Key: HUDI-8136 > URL: https://issues.apache.org/jira/browse/HUDI-8136 > Project: Apache Hudi > Issue Type: Improvement > Components: flink >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 1.0.0 > > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Created] (HUDI-8136) New instant time generation for Flink streaming pipeline
Danny Chen created HUDI-8136: Summary: New instant time generation for Flink streaming pipeline Key: HUDI-8136 URL: https://issues.apache.org/jira/browse/HUDI-8136 Project: Apache Hudi Issue Type: Improvement Components: flink Reporter: Danny Chen Fix For: 1.0.0 -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Updated] (HUDI-8136) New instant time generation for Flink streaming pipeline
[ https://issues.apache.org/jira/browse/HUDI-8136?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-8136: - Description: h2. Design The new design moves the instant time generation from being blocked on the checkpoint completion on the co-ordinator, to the writers cheaply obtaining the instant time to use before every file write by sending a light-weight request to the coordinator. The coordinator will maintain a mapping from checkpoint barriers(ids) to the instant times to be used by writers for file written during a checkpoint. When the writer sends a new instant time request, coordinator will return an existing pending instant or a new one(based on the comparison of the last finished checkpoint barrier from the writer and the existing barriers on the coordinator). We allow at most 2 pending instants on the coordinator. When an instant is committed, the pending mapping item from memory should be removed. Note that new instant time generation/switching on checkpoint start and instant time serving should happen in sequence (the same thread or a in-process lock) in the coordinator. The design assumes the following invariants.. * Once a writer task starts writing a file with time {{tx}} , it will not write any file with time {{ty}} , such that {{ty < tx}} * Checkpoint {{i}} will be started on the co-ordinator only after checkpoint {{j}} completes such that {{i < j}} Overall flow is as follows : # During startup co-ordinator generates a new instant {{tx}} and request it on the timeline. # ## Happens within a process-local lock shared by instant time generation ## Any request for instant times from writer tasks, will serve {{tx}} # Writer tasks fetch an instant time to use for any file written, by issuing a call to co-ordinator ## writer tasks keep track of all files written by them between checkpoints. # When starting a checkpoint, co-ordinator takes the lock again and generates a new instant {{ty}} , request it on the timeline ## This ensure all instant time fetches for {{tx}} are first served, before checkpoint starts. i.e they will be reported back to the co-ordinator when the checkpoint completes. # When they receive the event to checkpoint, the writer tasks flush all open files, return the list of files written as a part of the checkpoint. # Once co-ordinator receives a list of all files from all writer tasks, it can include two types of files. ## files belonging to {{tx}} , which can now be comitted since we know no files with time {{tx}} could have been written between steps 3 & 5. ## files belonging to {{ty}} , which cannot be committed yet (there could be files still being written for {{ty}} ), but co-ordinator needs to ensure these files are ultimately included when {{ty}} is committed. h4. The rollback of failed instants The clean policy should be always configured as lazy, when a {{ck_n, instant_m}} was committed to Hudi timeline, all the instants that are earlier than {{instant_m}} should invoke explicitly close to shutdown the heartbeat thread so that the async cleaner would finally roll back it. When a full task failover triggers, all the pending instants should be rolled back, but we would hand over this task to async cleaner. was: h2. Design The new design moves the instant time generation from being blocked on the checkpoint completion on the co-ordinator, to the writers cheaply obtaining the instant time to use before every file write by sending a light-weight request to the coordinator. The coordinator will maintain a mapping from checkpoint barriers(ids) to the instant times to be used by writers for file written during a checkpoint. When the writer sends a new instant time request, coordinator will return an existing pending instant or a new one(based on the comparison of the last finished checkpoint barrier from the writer and the existing barriers on the coordinator). We allow at most 2 pending instants on the coordinator. When an instant is committed, the pending mapping item from memory should be removed. Note that new instant time generation/switching on checkpoint start and instant time serving should happen in sequence (the same thread or a in-process lock) in the coordinator. The design assumes the following invariants.. * Once a writer task starts writing a file with time {{tx}} , it will not write any file with time {{ty}} , such that {{ty < tx}} * Checkpoint {{i}} will be started on the co-ordinator only after checkpoint {{j}} completes such that {{i < j}} Overall flow is as follows : # During startup co-ordinator generates a new instant {{tx}} and request it on the timeline. ## Happens within a process-local lock shared by instant time generation ## Any request for instant times from writer tasks, will serve {{tx}} # Writer tasks fetch an instant time to use for any file written, by issuing a call to co-ordinator ## wr
[jira] [Updated] (HUDI-8136) New instant time generation for Flink streaming pipeline
[ https://issues.apache.org/jira/browse/HUDI-8136?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-8136: - Description: h2. Design The new design moves the instant time generation from being blocked on the checkpoint completion on the co-ordinator, to the writers cheaply obtaining the instant time to use before every file write by sending a light-weight request to the coordinator. The coordinator will maintain a mapping from checkpoint barriers(ids) to the instant times to be used by writers for file written during a checkpoint. When the writer sends a new instant time request, coordinator will return an existing pending instant or a new one(based on the comparison of the last finished checkpoint barrier from the writer and the existing barriers on the coordinator). We allow at most 2 pending instants on the coordinator. When an instant is committed, the pending mapping item from memory should be removed. Note that new instant time generation/switching on checkpoint start and instant time serving should happen in sequence (the same thread or a in-process lock) in the coordinator. The design assumes the following invariants.. * Once a writer task starts writing a file with time {{tx}} , it will not write any file with time {{ty}} , such that {{ty < tx}} * Checkpoint {{i}} will be started on the co-ordinator only after checkpoint {{j}} completes such that {{i < j}} Overall flow is as follows : # During startup co-ordinator generates a new instant {{tx}} and request it on the timeline. ## Happens within a process-local lock shared by instant time generation ## Any request for instant times from writer tasks, will serve {{tx}} # Writer tasks fetch an instant time to use for any file written, by issuing a call to co-ordinator ## writer tasks keep track of all files written by them between checkpoints. # When starting a checkpoint, co-ordinator takes the lock again and generates a new instant {{ty}} , request it on the timeline ## This ensure all instant time fetches for {{tx}} are first served, before checkpoint starts. i.e they will be reported back to the co-ordinator when the checkpoint completes. # When they receive the event to checkpoint, the writer tasks flush all open files, return the list of files written as a part of the checkpoint. # Once co-ordinator receives a list of all files from all writer tasks, it can include two types of files. ## files belonging to {{tx}} , which can now be comitted since we know no files with time {{tx}} could have been written between steps 3 & 5. ## files belonging to {{ty}} , which cannot be committed yet (there could be files still being written for {{ty}} ), but co-ordinator needs to ensure these files are ultimately included when {{ty}} is committed. > New instant time generation for Flink streaming pipeline > > > Key: HUDI-8136 > URL: https://issues.apache.org/jira/browse/HUDI-8136 > Project: Apache Hudi > Issue Type: Improvement > Components: flink >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 1.0.0 > > > h2. Design > The new design moves the instant time generation from being blocked on the > checkpoint completion on the co-ordinator, to the writers cheaply obtaining > the instant time to use before every file write by sending a light-weight > request to the coordinator. The coordinator will maintain a mapping from > checkpoint barriers(ids) to the instant times to be used by writers for file > written during a checkpoint. When the writer sends a new instant time > request, coordinator will return an existing pending instant or a new > one(based on the comparison of the last finished checkpoint barrier from the > writer and the existing barriers on the coordinator). We allow at most 2 > pending instants on the coordinator. When an instant is committed, the > pending mapping item from memory should be removed. Note that new instant > time generation/switching on checkpoint start and instant time serving should > happen in sequence (the same thread or a in-process lock) in the coordinator. > The design assumes the following invariants.. * Once a writer task starts > writing a file with time {{tx}} , it will not write any file with time {{ty}} > , such that {{ty < tx}} > * Checkpoint {{i}} will be started on the co-ordinator only after checkpoint > {{j}} completes such that {{i < j}} > > Overall flow is as follows : # During startup co-ordinator generates a new > instant {{tx}} and request it on the timeline. > ## Happens within a process-local lock shared by instant time generation > ## Any request for instant times from writer tasks, will serve {{tx}} > # Writer tasks fetch an instant time to use for any file written, by issu
[jira] [Created] (HUDI-2143) Tweak the default compaction target IO to 500GB when flink async compaction is off
Danny Chen created HUDI-2143: Summary: Tweak the default compaction target IO to 500GB when flink async compaction is off Key: HUDI-2143 URL: https://issues.apache.org/jira/browse/HUDI-2143 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HUDI-2162) Instant is null cause flushBuffer failed in casual
[ https://issues.apache.org/jira/browse/HUDI-2162?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17378949#comment-17378949 ] Danny Chen commented on HUDI-2162: -- You should set up the timeout correctly. > Instant is null cause flushBuffer failed in casual > -- > > Key: HUDI-2162 > URL: https://issues.apache.org/jira/browse/HUDI-2162 > Project: Apache Hudi > Issue Type: Bug > Components: Flink Integration >Reporter: zhangminglei >Assignee: zhangminglei >Priority: Blocker > > Since commit Instant and getting Instant are asynchronous , and thus instant > is null the default waiting time is 0 must greater than ckpTimeout would > cause Exception belows as shown. > WRITE_COMMIT_ACK_TIMEOUT is for internal usage. so it is not suitable for > java api user under exactly once, This kind of usage is too weak under the > context. > Timeout(0ms) while waiting for instant null to commit > at > org.apache.hudi.sink.StreamWriteFunction.instantToWrite(StreamWriteFunction.java:597) > at > org.apache.hudi.sink.StreamWriteFunction.flushBucket(StreamWriteFunction.java:618) > at > org.apache.hudi.sink.StreamWriteFunction.bufferRecord(StreamWriteFunction.java:554) > at > org.apache.hudi.sink.StreamWriteFunction.processElement(StreamWriteFunction.java:236) > at > org.apache.flink.streaming.api.operators.ProcessOperator.processElement(ProcessOperator.java:66) -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2170) Always choose the latest record for HoodieRecordPayload
Danny Chen created HUDI-2170: Summary: Always choose the latest record for HoodieRecordPayload Key: HUDI-2170 URL: https://issues.apache.org/jira/browse/HUDI-2170 Project: Apache Hudi Issue Type: Improvement Components: Common Core Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Now in {{OverwriteWithLatestAvroPayload.preCombine}}, we still choose the old record when the new record has the same preCombine field with the old one, actually it is more natural to keep the new incoming record instead. The {{DefaultHoodieRecordPayload.combineAndGetUpdateValue}} method already did that. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Updated] (HUDI-2170) Always choose the latest record for HoodieRecordPayload
[ https://issues.apache.org/jira/browse/HUDI-2170?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-2170: - Description: Now in {{OverwriteWithLatestAvroPayload.preCombine}}, we still choose the old record when the new record has the same preCombine field with the old one, actually it is more natural to keep the new incoming record instead. The {{DefaultHoodieRecordPayload.combineAndGetUpdateValue}} method already did that. See issue: https://github.com/apache/hudi/issues/3266. was:Now in {{OverwriteWithLatestAvroPayload.preCombine}}, we still choose the old record when the new record has the same preCombine field with the old one, actually it is more natural to keep the new incoming record instead. The {{DefaultHoodieRecordPayload.combineAndGetUpdateValue}} method already did that. > Always choose the latest record for HoodieRecordPayload > --- > > Key: HUDI-2170 > URL: https://issues.apache.org/jira/browse/HUDI-2170 > Project: Apache Hudi > Issue Type: Improvement > Components: Common Core >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > > Now in {{OverwriteWithLatestAvroPayload.preCombine}}, we still choose the old > record when the new record has the same preCombine field with the old one, > actually it is more natural to keep the new incoming record instead. The > {{DefaultHoodieRecordPayload.combineAndGetUpdateValue}} method already did > that. > See issue: https://github.com/apache/hudi/issues/3266. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2171) Add parallelism conf for bootstrap operator
[ https://issues.apache.org/jira/browse/HUDI-2171?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2171. -- Fix Version/s: 0.9.0 Resolution: Fixed Fixed via master branch: 632bfd1a65f55deff60bd56e514738b9c8730140 > Add parallelism conf for bootstrap operator > --- > > Key: HUDI-2171 > URL: https://issues.apache.org/jira/browse/HUDI-2171 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: yuzhaojing >Assignee: yuzhaojing >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > > Add parallelism conf for bootstrap operator -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2153) BucketAssignFunction NullPointerException
[ https://issues.apache.org/jira/browse/HUDI-2153?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2153. -- Assignee: Danny Chen Resolution: Fixed Fixed via master branch: 23a4a96eb416c13d8b1f921cc51286f93d61d9b3 > BucketAssignFunction NullPointerException > - > > Key: HUDI-2153 > URL: https://issues.apache.org/jira/browse/HUDI-2153 > Project: Apache Hudi > Issue Type: Bug >Reporter: moran >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > > java.lang.NullPointerException > at > org.apache.hudi.sink.partitioner.BucketAssignFunction.processRecord(BucketAssignFunction.java:198) > at > org.apache.hudi.sink.partitioner.BucketAssignFunction.processElement(BucketAssignFunction.java:159) > at > org.apache.flink.streaming.api.operators.KeyedProcessOperator.processElement(KeyedProcessOperator.java:83) > at > org.apache.flink.streaming.runtime.tasks.OneInputStreamTask$StreamTaskNetworkOutput.emitRecord(OneInputStreamTask.java:191) > at > org.apache.flink.streaming.runtime.io.StreamTaskNetworkInput.processElement(StreamTaskNetworkInput.java:204) > at > org.apache.flink.streaming.runtime.io.StreamTaskNetworkInput.emitNext(StreamTaskNetworkInput.java:174) > at > org.apache.flink.streaming.runtime.io.StreamOneInputProcessor.processInput(StreamOneInputProcessor.java:65) > at > org.apache.flink.streaming.runtime.tasks.StreamTask.processInput(StreamTask.java:396) > at > org.apache.flink.streaming.runtime.tasks.mailbox.MailboxProcessor.runMailboxLoop(MailboxProcessor.java:191) > at > org.apache.flink.streaming.runtime.tasks.StreamTask.runMailboxLoop(StreamTask.java:617) > at > org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:581) > at org.apache.flink.runtime.taskmanager.Task.doRun(Task.java:755) > at org.apache.flink.runtime.taskmanager.Task.run(Task.java:570) > at java.lang.Thread.run(Thread.java:748) > ERROR at > Line 197 of the BucketAssignFunction class > (this.context.setCurrentKey(recordKey)) > Why is this context null -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HUDI-2181) Refine doc for FlinkCreateHandle
[ https://issues.apache.org/jira/browse/HUDI-2181?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17381287#comment-17381287 ] Danny Chen commented on HUDI-2181: -- Reasonable, welcome for the contribution ~ > Refine doc for FlinkCreateHandle > > > Key: HUDI-2181 > URL: https://issues.apache.org/jira/browse/HUDI-2181 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: zhangminglei >Assignee: zhangminglei >Priority: Major > > FlinkCreateHandle does not append to the original file for subsequent > mini-batches, instead every inserts batch would create a new file. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2185) Remove the default parallelism of index bootstrap and bucket assigner
Danny Chen created HUDI-2185: Summary: Remove the default parallelism of index bootstrap and bucket assigner Key: HUDI-2185 URL: https://issues.apache.org/jira/browse/HUDI-2185 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2185) Remove the default parallelism of index bootstrap and bucket assigner
[ https://issues.apache.org/jira/browse/HUDI-2185?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2185. -- Resolution: Fixed Fixed via master branch: c8aaf00819b7af9c5575662868aa40f702a8aa1e > Remove the default parallelism of index bootstrap and bucket assigner > - > > Key: HUDI-2185 > URL: https://issues.apache.org/jira/browse/HUDI-2185 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Reopened] (HUDI-2087) Support Append only in Flink stream
[ https://issues.apache.org/jira/browse/HUDI-2087?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen reopened HUDI-2087: -- Revert and reopen for more disscussion. > Support Append only in Flink stream > --- > > Key: HUDI-2087 > URL: https://issues.apache.org/jira/browse/HUDI-2087 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: yuzhaojing >Assignee: yuzhaojing >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > Attachments: image-2021-07-08-22-04-30-039.png, > image-2021-07-08-22-04-40-018.png > > > It is necessary to support append mode in flink stream, as the data lake > should be able to write log type data as parquet high performance without > merge. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2191) Bump flink version to 1.13.1
Danny Chen created HUDI-2191: Summary: Bump flink version to 1.13.1 Key: HUDI-2191 URL: https://issues.apache.org/jira/browse/HUDI-2191 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Aims to use flink 1.13.1 for 0.9.0 release. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2193) Remove state in BootstrapFunction
[ https://issues.apache.org/jira/browse/HUDI-2193?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2193. -- Fix Version/s: 0.9.0 Resolution: Fixed Fixed via master branch: 2099bf41db76e9a6e946aa41c318b7c0e18be04d > Remove state in BootstrapFunction > - > > Key: HUDI-2193 > URL: https://issues.apache.org/jira/browse/HUDI-2193 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: yuzhaojing >Assignee: yuzhaojing >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > > Remove state in BootstrapFunction to support restart job with out bootstrap -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2198) Clean and reset the bootstrap events for coordinator when task failover
Danny Chen created HUDI-2198: Summary: Clean and reset the bootstrap events for coordinator when task failover Key: HUDI-2198 URL: https://issues.apache.org/jira/browse/HUDI-2198 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2145) Create new bucket when NewFileAssignState filled
[ https://issues.apache.org/jira/browse/HUDI-2145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2145. -- Fix Version/s: 0.9.0 Resolution: Fixed Fixed via master branch: 634163a990569aa4463b58830396f455dd15340c > Create new bucket when NewFileAssignState filled > > > Key: HUDI-2145 > URL: https://issues.apache.org/jira/browse/HUDI-2145 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: yuzhaojing >Assignee: yuzhaojing >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2198) Clean and reset the bootstrap events for coordinator when task failover
[ https://issues.apache.org/jira/browse/HUDI-2198?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2198. -- Fix Version/s: 0.9.0 Resolution: Fixed Fixed via master branch: 858e84b5b2f1c5fbd4266922b494ad0d16f5b92a > Clean and reset the bootstrap events for coordinator when task failover > --- > > Key: HUDI-2198 > URL: https://issues.apache.org/jira/browse/HUDI-2198 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2204) Add marker files for flink writer
Danny Chen created HUDI-2204: Summary: Add marker files for flink writer Key: HUDI-2204 URL: https://issues.apache.org/jira/browse/HUDI-2204 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2205) Rollback inflight compaction for flink writer
Danny Chen created HUDI-2205: Summary: Rollback inflight compaction for flink writer Key: HUDI-2205 URL: https://issues.apache.org/jira/browse/HUDI-2205 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2204) Add marker files for flink writer
[ https://issues.apache.org/jira/browse/HUDI-2204?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2204. -- Resolution: Fixed Fixed via master branch: 2370a9facbe4418f994f29c426e9b2a255e3abb0 > Add marker files for flink writer > - > > Key: HUDI-2204 > URL: https://issues.apache.org/jira/browse/HUDI-2204 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2206) Fix checkpoint blocked because getLastPendingInstant() action after than restoreWriteMetadata() action
[ https://issues.apache.org/jira/browse/HUDI-2206?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2206. -- Assignee: Danny Chen Resolution: Fixed Fixed via master branch: fe5d2e7f53eb36d9e35acd8d944e59d988dba475 > Fix checkpoint blocked because getLastPendingInstant() action after than > restoreWriteMetadata() action > --- > > Key: HUDI-2206 > URL: https://issues.apache.org/jira/browse/HUDI-2206 > Project: Apache Hudi > Issue Type: Bug > Components: Flink Integration >Reporter: Zheng yunhong >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > > Fix checkpoint blocked because getLastPendingInstant() action after than > restoreWriteMetadata() action which will cause a deadlock during checkpoint. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2209) Bulk insert for flink writer
Danny Chen created HUDI-2209: Summary: Bulk insert for flink writer Key: HUDI-2209 URL: https://issues.apache.org/jira/browse/HUDI-2209 Project: Apache Hudi Issue Type: New Feature Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2219) Fix NPE of HoodieConfig
Danny Chen created HUDI-2219: Summary: Fix NPE of HoodieConfig Key: HUDI-2219 URL: https://issues.apache.org/jira/browse/HUDI-2219 Project: Apache Hudi Issue Type: Bug Components: Compaction Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2209) Bulk insert for flink writer
[ https://issues.apache.org/jira/browse/HUDI-2209?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2209. -- Resolution: Fixed Fixed via master branch: 9d2a65a6a6ff9add81411147f1cddd03f7c08e6c > Bulk insert for flink writer > > > Key: HUDI-2209 > URL: https://issues.apache.org/jira/browse/HUDI-2209 > Project: Apache Hudi > Issue Type: New Feature > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2227) Only sync hive meta on successful commit for flink batch writer
Danny Chen created HUDI-2227: Summary: Only sync hive meta on successful commit for flink batch writer Key: HUDI-2227 URL: https://issues.apache.org/jira/browse/HUDI-2227 Project: Apache Hudi Issue Type: Bug Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2219) Fix NPE of HoodieConfig
[ https://issues.apache.org/jira/browse/HUDI-2219?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2219. -- Resolution: Fixed Fixed via master branch: ab2e0d0ba2697ba0750ec52fbb3b3a0187734a4b > Fix NPE of HoodieConfig > --- > > Key: HUDI-2219 > URL: https://issues.apache.org/jira/browse/HUDI-2219 > Project: Apache Hudi > Issue Type: Bug > Components: Compaction >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2227) Only sync hive meta on successful commit for flink batch writer
[ https://issues.apache.org/jira/browse/HUDI-2227?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2227. -- Resolution: Fixed Fixed via master branch: 60758b36ea033a5985eabc53da5f6036ffaa5c0d > Only sync hive meta on successful commit for flink batch writer > --- > > Key: HUDI-2227 > URL: https://issues.apache.org/jira/browse/HUDI-2227 > Project: Apache Hudi > Issue Type: Bug > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2215) Add rateLimiter when Flink writes to hudi
[ https://issues.apache.org/jira/browse/HUDI-2215?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2215. -- Resolution: Fixed Fixed via master branch: 00cd35f90ad0cd96a943edd148bf881fc4b7bb5b > Add rateLimiter when Flink writes to hudi > - > > Key: HUDI-2215 > URL: https://issues.apache.org/jira/browse/HUDI-2215 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: WangMinChao >Assignee: WangMinChao >Priority: Minor > Labels: flink, pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2241) Explicit parallelism for flink bulk insert
Danny Chen created HUDI-2241: Summary: Explicit parallelism for flink bulk insert Key: HUDI-2241 URL: https://issues.apache.org/jira/browse/HUDI-2241 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2245) BucketAssigner generates the fileId evenly to avoid data skew
Danny Chen created HUDI-2245: Summary: BucketAssigner generates the fileId evenly to avoid data skew Key: HUDI-2245 URL: https://issues.apache.org/jira/browse/HUDI-2245 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2246) Time travel query for flink sql
Danny Chen created HUDI-2246: Summary: Time travel query for flink sql Key: HUDI-2246 URL: https://issues.apache.org/jira/browse/HUDI-2246 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2245) BucketAssigner generates the fileId evenly to avoid data skew
[ https://issues.apache.org/jira/browse/HUDI-2245?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2245. -- Resolution: Fixed Fixed via master branch: 91c221341293e80c28cce19f1642199495a96f66 > BucketAssigner generates the fileId evenly to avoid data skew > - > > Key: HUDI-2245 > URL: https://issues.apache.org/jira/browse/HUDI-2245 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2228) Add option 'hive_sync.mode' for flink writer
[ https://issues.apache.org/jira/browse/HUDI-2228?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2228. -- Resolution: Fixed Fixed via master branch: 7739518879cf1c30576b938d705947192a2dd9ad > Add option 'hive_sync.mode' for flink writer > > > Key: HUDI-2228 > URL: https://issues.apache.org/jira/browse/HUDI-2228 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Zheng yunhong >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > > [Add option 'hive_sync.mode' for flink > writer|https://github.com/apache/hudi/pull/3352/commits/9ff4b7382c22745aed8dfc96e93bdc53fbd51a25]. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2241) Explicit parallelism for flink bulk insert
[ https://issues.apache.org/jira/browse/HUDI-2241?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2241. -- Resolution: Fixed Fixed via master branch: efbbb67420b7082a3960f3d32215dabd959f5525 > Explicit parallelism for flink bulk insert > -- > > Key: HUDI-2241 > URL: https://issues.apache.org/jira/browse/HUDI-2241 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2254) Builtin sort operator for flink bulk insert
Danny Chen created HUDI-2254: Summary: Builtin sort operator for flink bulk insert Key: HUDI-2254 URL: https://issues.apache.org/jira/browse/HUDI-2254 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Updated] (HUDI-2252) Default consumes from the latest instant for flink streaming reader
[ https://issues.apache.org/jira/browse/HUDI-2252?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-2252: - Summary: Default consumes from the latest instant for flink streaming reader (was: Replace read full data with read latest commit data in flink stream read) > Default consumes from the latest instant for flink streaming reader > --- > > Key: HUDI-2252 > URL: https://issues.apache.org/jira/browse/HUDI-2252 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Zheng yunhong >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > > Replace read full data with read latest commit data in flink stream read. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2252) Replace read full data with read latest commit data in flink stream read
[ https://issues.apache.org/jira/browse/HUDI-2252?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2252. -- Resolution: Fixed Fixed via master branch: 8b19ec9ca07070a9819502e82091dd14d559ef94 > Replace read full data with read latest commit data in flink stream read > - > > Key: HUDI-2252 > URL: https://issues.apache.org/jira/browse/HUDI-2252 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Zheng yunhong >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > > Replace read full data with read latest commit data in flink stream read. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1949) Refactor BucketAssigner to make it more efficient
Danny Chen created HUDI-1949: Summary: Refactor BucketAssigner to make it more efficient Key: HUDI-1949 URL: https://issues.apache.org/jira/browse/HUDI-1949 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Add a process single class {{WriteProfile}}, the record and small files profile re-construction can be more efficient if we reuse by same checkpoint id. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1952) Support hive3 meta sync for flink writer
Danny Chen created HUDI-1952: Summary: Support hive3 meta sync for flink writer Key: HUDI-1952 URL: https://issues.apache.org/jira/browse/HUDI-1952 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1956) BucketAssignFunction use ValueState instead of MapState
Danny Chen created HUDI-1956: Summary: BucketAssignFunction use ValueState instead of MapState Key: HUDI-1956 URL: https://issues.apache.org/jira/browse/HUDI-1956 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Use the value state to reduce the memory footprint. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Closed] (HUDI-1956) BucketAssignFunction use ValueState instead of MapState
[ https://issues.apache.org/jira/browse/HUDI-1956?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-1956. Resolution: Duplicate > BucketAssignFunction use ValueState instead of MapState > --- > > Key: HUDI-1956 > URL: https://issues.apache.org/jira/browse/HUDI-1956 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > > Use the value state to reduce the memory footprint. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1961) Add a debezium json integration test case for flink
Danny Chen created HUDI-1961: Summary: Add a debezium json integration test case for flink Key: HUDI-1961 URL: https://issues.apache.org/jira/browse/HUDI-1961 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1967) Fix the NPE for MOR Hive rt table query
Danny Chen created HUDI-1967: Summary: Fix the NPE for MOR Hive rt table query Key: HUDI-1967 URL: https://issues.apache.org/jira/browse/HUDI-1967 Project: Apache Hudi Issue Type: Bug Components: Hive Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Updated] (HUDI-1967) Fix the NPE for MOR Hive rt table query
[ https://issues.apache.org/jira/browse/HUDI-1967?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-1967: - Description: See the discussion here: https://github.com/apache/hudi/issues/2874 > Fix the NPE for MOR Hive rt table query > > > Key: HUDI-1967 > URL: https://issues.apache.org/jira/browse/HUDI-1967 > Project: Apache Hudi > Issue Type: Bug > Components: Hive Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > > See the discussion here: https://github.com/apache/hudi/issues/2874 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1969) Support reading logs for MOR Hive rt table
Danny Chen created HUDI-1969: Summary: Support reading logs for MOR Hive rt table Key: HUDI-1969 URL: https://issues.apache.org/jira/browse/HUDI-1969 Project: Apache Hudi Issue Type: Improvement Components: Hive Integration Reporter: Danny Chen -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-1756) Assigns the buckets by record key for Flink writer
[ https://issues.apache.org/jira/browse/HUDI-1756?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-1756. -- Resolution: Duplicate > Assigns the buckets by record key for Flink writer > -- > > Key: HUDI-1756 > URL: https://issues.apache.org/jira/browse/HUDI-1756 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Closed] (HUDI-1819) Remove legacy code for Flink writer
[ https://issues.apache.org/jira/browse/HUDI-1819?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-1819. Assignee: Danny Chen Resolution: Duplicate > Remove legacy code for Flink writer > --- > > Key: HUDI-1819 > URL: https://issues.apache.org/jira/browse/HUDI-1819 > Project: Apache Hudi > Issue Type: Task > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > > Removes the useless code to avoid confusion. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Closed] (HUDI-1857) Shade google guava for hudi-flink-bundle jar
[ https://issues.apache.org/jira/browse/HUDI-1857?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-1857. Resolution: Fixed > Shade google guava for hudi-flink-bundle jar > > > Key: HUDI-1857 > URL: https://issues.apache.org/jira/browse/HUDI-1857 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > > Shade out the google guava to avoid conflicts. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1986) Skip creating marker files for flink merge handle
Danny Chen created HUDI-1986: Summary: Skip creating marker files for flink merge handle Key: HUDI-1986 URL: https://issues.apache.org/jira/browse/HUDI-1986 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Skip creating the marker files for flink merge handle to make it more robust. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1987) Fix non partition table hive meta sync for flink writer
Danny Chen created HUDI-1987: Summary: Fix non partition table hive meta sync for flink writer Key: HUDI-1987 URL: https://issues.apache.org/jira/browse/HUDI-1987 Project: Apache Hudi Issue Type: Bug Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1992) Release the new records map for merge handle #close
Danny Chen created HUDI-1992: Summary: Release the new records map for merge handle #close Key: HUDI-1992 URL: https://issues.apache.org/jira/browse/HUDI-1992 Project: Apache Hudi Issue Type: Improvement Components: Common Core Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 The new records map can be cleaned eagerly to save the memory footprint. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1994) [HUDI-1992] Release the new records iterator for append handle #close
Danny Chen created HUDI-1994: Summary: [HUDI-1992] Release the new records iterator for append handle #close Key: HUDI-1994 URL: https://issues.apache.org/jira/browse/HUDI-1994 Project: Apache Hudi Issue Type: Improvement Components: Common Core Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 To save the memory footprint. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Updated] (HUDI-1994) Release the new records iterator for append handle #close
[ https://issues.apache.org/jira/browse/HUDI-1994?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-1994: - Summary: Release the new records iterator for append handle #close (was: [HUDI-1992] Release the new records iterator for append handle #close) > Release the new records iterator for append handle #close > - > > Key: HUDI-1994 > URL: https://issues.apache.org/jira/browse/HUDI-1994 > Project: Apache Hudi > Issue Type: Improvement > Components: Common Core >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > > To save the memory footprint. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-1999) Refresh the base file view cache for WriteProfile
Danny Chen created HUDI-1999: Summary: Refresh the base file view cache for WriteProfile Key: HUDI-1999 URL: https://issues.apache.org/jira/browse/HUDI-1999 Project: Apache Hudi Issue Type: Bug Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Refresh the view to discover new small files. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2015) Fix flink operator uid to allow multiple pipelines in one job
Danny Chen created HUDI-2015: Summary: Fix flink operator uid to allow multiple pipelines in one job Key: HUDI-2015 URL: https://issues.apache.org/jira/browse/HUDI-2015 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Fix the uid conflicts. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2030) Add metadata cache to WriteProfile to reduce IO
Danny Chen created HUDI-2030: Summary: Add metadata cache to WriteProfile to reduce IO Key: HUDI-2030 URL: https://issues.apache.org/jira/browse/HUDI-2030 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Keeps same number of instant metadata cache and refresh the cache on new commits. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2036) Move the compaction plan scheduling out of flink writer coordinator
Danny Chen created HUDI-2036: Summary: Move the compaction plan scheduling out of flink writer coordinator Key: HUDI-2036 URL: https://issues.apache.org/jira/browse/HUDI-2036 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Since HUDI-1955 was fixed, we can move the scheduling out if the coordinator to make the coordinator more lightweight. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2037) Move the compaction plan scheduling out of flink writer coordinator
Danny Chen created HUDI-2037: Summary: Move the compaction plan scheduling out of flink writer coordinator Key: HUDI-2037 URL: https://issues.apache.org/jira/browse/HUDI-2037 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 Since HUDI-1955 was fixed, we can move the scheduling out if the coordinator to make the coordinator more lightweight. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Closed] (HUDI-2037) Move the compaction plan scheduling out of flink writer coordinator
[ https://issues.apache.org/jira/browse/HUDI-2037?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen closed HUDI-2037. Resolution: Invalid > Move the compaction plan scheduling out of flink writer coordinator > --- > > Key: HUDI-2037 > URL: https://issues.apache.org/jira/browse/HUDI-2037 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > > Since HUDI-1955 was fixed, we can move the scheduling out if the coordinator > to make the coordinator more lightweight. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2040) Make flink writer as exactly-once by default
Danny Chen created HUDI-2040: Summary: Make flink writer as exactly-once by default Key: HUDI-2040 URL: https://issues.apache.org/jira/browse/HUDI-2040 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2048) HoodieRealtimeInputFormatUtils#groupLogsByBaseFile throws NPE for file group that has only logs
Danny Chen created HUDI-2048: Summary: HoodieRealtimeInputFormatUtils#groupLogsByBaseFile throws NPE for file group that has only logs Key: HUDI-2048 URL: https://issues.apache.org/jira/browse/HUDI-2048 Project: Apache Hudi Issue Type: Bug Components: Hive Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Assigned] (HUDI-2049) StreamWriteFunction should wait for the next inflight instant time before flushing
[ https://issues.apache.org/jira/browse/HUDI-2049?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen reassigned HUDI-2049: Assignee: Danny Chen > StreamWriteFunction should wait for the next inflight instant time before > flushing > --- > > Key: HUDI-2049 > URL: https://issues.apache.org/jira/browse/HUDI-2049 > Project: Apache Hudi > Issue Type: Bug > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2049) StreamWriteFunction should wait for the next inflight instant time before flushing
Danny Chen created HUDI-2049: Summary: StreamWriteFunction should wait for the next inflight instant time before flushing Key: HUDI-2049 URL: https://issues.apache.org/jira/browse/HUDI-2049 Project: Apache Hudi Issue Type: Bug Components: Flink Integration Reporter: Danny Chen -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Updated] (HUDI-2049) StreamWriteFunction should wait for the next inflight instant time before flushing
[ https://issues.apache.org/jira/browse/HUDI-2049?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-2049: - Fix Version/s: 0.9.0 > StreamWriteFunction should wait for the next inflight instant time before > flushing > --- > > Key: HUDI-2049 > URL: https://issues.apache.org/jira/browse/HUDI-2049 > Project: Apache Hudi > Issue Type: Bug > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2054) Remove the duplicate name for flink write pipeline
Danny Chen created HUDI-2054: Summary: Remove the duplicate name for flink write pipeline Key: HUDI-2054 URL: https://issues.apache.org/jira/browse/HUDI-2054 Project: Apache Hudi Issue Type: Task Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2067) Sync all the options of FlinkOptions to FlinkStreamerConfig
Danny Chen created HUDI-2067: Summary: Sync all the options of FlinkOptions to FlinkStreamerConfig Key: HUDI-2067 URL: https://issues.apache.org/jira/browse/HUDI-2067 Project: Apache Hudi Issue Type: Task Components: Flink Integration Reporter: Danny Chen Fix For: 0.9.0 Sync the options so that the {{HoodieFlinkStreamer}} can have more config options. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2068) Skip the assign state for SmallFileAssign when the state can not assign initially
Danny Chen created HUDI-2068: Summary: Skip the assign state for SmallFileAssign when the state can not assign initially Key: HUDI-2068 URL: https://issues.apache.org/jira/browse/HUDI-2068 Project: Apache Hudi Issue Type: Bug Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2074) Use while loop instead of recursive call in MergeOnReadInputFormat#MergeIterator to avoid StackOverflow
Danny Chen created HUDI-2074: Summary: Use while loop instead of recursive call in MergeOnReadInputFormat#MergeIterator to avoid StackOverflow Key: HUDI-2074 URL: https://issues.apache.org/jira/browse/HUDI-2074 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2084) Resend the uncommitted write metadata when start up
Danny Chen created HUDI-2084: Summary: Resend the uncommitted write metadata when start up Key: HUDI-2084 URL: https://issues.apache.org/jira/browse/HUDI-2084 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Resolved] (HUDI-2084) Resend the uncommitted write metadata when start up
[ https://issues.apache.org/jira/browse/HUDI-2084?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen resolved HUDI-2084. -- Resolution: Fixed Resolved via: 37b7c65d8a3ede00ae16909a06e31c24f179998c > Resend the uncommitted write metadata when start up > --- > > Key: HUDI-2084 > URL: https://issues.apache.org/jira/browse/HUDI-2084 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Labels: pull-request-available > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2094) Supports hive style partitioning for flink writer
Danny Chen created HUDI-2094: Summary: Supports hive style partitioning for flink writer Key: HUDI-2094 URL: https://issues.apache.org/jira/browse/HUDI-2094 Project: Apache Hudi Issue Type: New Feature Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2112) Support reading pure logs file group for flink batch reader
Danny Chen created HUDI-2112: Summary: Support reading pure logs file group for flink batch reader Key: HUDI-2112 URL: https://issues.apache.org/jira/browse/HUDI-2112 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Updated] (HUDI-2112) Support reading pure logs file group for flink batch reader after compaction
[ https://issues.apache.org/jira/browse/HUDI-2112?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Danny Chen updated HUDI-2112: - Summary: Support reading pure logs file group for flink batch reader after compaction (was: Support reading pure logs file group for flink batch reader) > Support reading pure logs file group for flink batch reader after compaction > > > Key: HUDI-2112 > URL: https://issues.apache.org/jira/browse/HUDI-2112 > Project: Apache Hudi > Issue Type: Improvement > Components: Flink Integration >Reporter: Danny Chen >Assignee: Danny Chen >Priority: Major > Fix For: 0.9.0 > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2121) Add operator uid for flink stateful operators
Danny Chen created HUDI-2121: Summary: Add operator uid for flink stateful operators Key: HUDI-2121 URL: https://issues.apache.org/jira/browse/HUDI-2121 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2126) The coordinator send events to write function when there are no data for the checkpoint
Danny Chen created HUDI-2126: Summary: The coordinator send events to write function when there are no data for the checkpoint Key: HUDI-2126 URL: https://issues.apache.org/jira/browse/HUDI-2126 Project: Apache Hudi Issue Type: Improvement Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (HUDI-2129) StreamerUtil.medianInstantTime should return a valid date time string
Danny Chen created HUDI-2129: Summary: StreamerUtil.medianInstantTime should return a valid date time string Key: HUDI-2129 URL: https://issues.apache.org/jira/browse/HUDI-2129 Project: Apache Hudi Issue Type: Bug Components: Flink Integration Reporter: Danny Chen Assignee: Danny Chen Fix For: 0.9.0 -- This message was sent by Atlassian Jira (v8.3.4#803005)