You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@iotdb.apache.org by Haonan Hou <ha...@apache.org> on 2026/07/02 08:46:25 UTC
Re: Proposal to Skip IoTDB 2.0.9 and Release 2.0.10 Directly
Hi,
I have a draft of 2.0.10 release notes. Let me list here.
Features & Improvements
- Data Query: Support schema-level and table-level storage space statistics
- Data Query: Add DEBUG SQL capability and optimize the result set of Explain Analyze
- Data Query: Support set operations (UNION/INTERSECT/EXCEPT) and Common Table Expressions (CTE) in the table model
- Data Query: Add IF scalar function, binary functions and APPROX_PERCENTILE aggregate function for the table model
- Data Query: Optimize display fields of the show queries command, add three new columns: client IP that initiates the query, timeout threshold, and total request waiting latency
- Storage Management: Provide C-language driver SDK interfaces, including parameter binding, automatic failover for multi-node switching and other fault-tolerant capabilities
- System Management: Add the capability to manually trigger node load balancing
- System Management: Add progress query capability for the remove datanode operation
- System Management: Introduce the show configuration statement to view cluster configuration information
- System Management: Add statistics for the number of slow write requests
- AI Management: Add two built-in models: Moirai2 and Toto
- AI Management: Enable AINode to manage multiple models under the same model_type
- AI Management: Support loading built-in models for CPU inference with isolated inference processes
Bugs
- Fix the issue that partition tables get lost when leader switch happens on single-node ConfigNode
- Fix the bug where configurations delivered via set configuration to a single DataNode take effect on all other DataNodes unexpectedly
- Fix the occasional continuous accumulation of reservedSize for timeout-only queries across multiple devices in tree model, which triggers query errors after hitting the max capacity limit
- Fix the parsing error in tree-model-to-table conversion logic when tree paths consist entirely of numeric characters
- Correct abnormal output of show devices root.db.** when root.db itself is a device, making it display root.db properly
- Fix the failure of kill query and query timeout mechanism when the client holds the connection open without fully consuming all result sets
- Fix array index out-of-bounds error caused by null values in attribute fields on the receiver side, which blocks the data synchronization pipeline
- Fix connection pool leakage exceptions in the C# client
- Fix exceptions occurring in the C++ client when write redirection is enabled while dn_rpc_address is configured to 0.0.0.0 on DataNode server
- Fix abnormal repeated broadcasting of deletion markers (Deletions) during the load operation
- Fix the issue that rows with null inserted fields cannot be synchronized to the receiver in real-time data synchronization scenarios
- Fix incorrect idempotency logic for out-of-order TTL-expired data on the synchronization receiver, which leads to repeated synchronization of expired data
- Fix the problem that enable-send-tsfile-limit parameter cannot be manually configured for historical Pipes split from full-data Pipe tasks
- Fix Pipe temporary stoppage triggered by insufficient memory of a single subtask within synchronization jobs
- Fix the issue that modifying username/password of source or sink in write-back-sink model generates error logs and prevents data from being received on the sink side
- Fix permission issue preventing non-root users from stopping the AINode process
- Mitigate memory leaks that tend to occur during long-running continuous inference of models running on CPU
BR,
Haonan
On 2026/06/30 09:38:46 Yuan Tian wrote:
> Hi all,
>
> We would like to skip the IoTDB 2.0.9 release and proceed directly with the
> 2.0.10 release.
>
> The main reason is that one release-note item originally planned for 2.0.9
> remained under active development and iteration longer than expected. By
> the time this item was fully completed, the development work planned for
> 2.0.10 had also been finished.
>
> Given this situation, releasing 2.0.9 separately would bring limited
> practical value and may instead cause unnecessary confusion. Therefore, we
> plan to include the relevant changes in 2.0.10 and make 2.0.10 the next
> release version.
>
> To make future release planning clearer, starting from 2.0.11, before each
> release iteration begins, we will send an email to the mailing list
> outlining the release-note items expected to be included in that version,
> together with the expected code-freeze date. Everyone is also welcome to
> propose additional release-note items that should be included in the
> corresponding version. Please note that the code-freeze date is not the
> same as the release date. After code freeze, we will still need to test and
> verify the corresponding release-note items before proceeding with the
> formal release.
>
> Please let us know if you have any concerns or objections.
>
> Best regards,
>
> ------------------------
>
> Yuan Tian
>
Re: Proposal to Skip IoTDB 2.0.9 and Release 2.0.10 Directly
Posted by Yuan Tian <ja...@gmail.com>.
LGTM
On Thu, Jul 2, 2026 at 5:11 PM Haonan Hou <ha...@apache.org> wrote:
> Hi Yuan,
>
> I have submit a PR [1] to do it. Please take a look.
> [1] https://github.com/apache/iotdb/pull/18093
>
> Best,
> Haonan
>
> On 2026/07/02 08:53:53 Yuan Tian wrote:
> > Hi haonan,
> >
> > LGTM, since rc/2.0.10 has already been checked out, I think we can bump
> > the version of master to 2.0.11-SNAPSHOT?
> >
> >
> > Best regards,
> > ---------------------
> > Yuan Tian
> >
> > On Thu, Jul 2, 2026 at 4:47 PM Haonan Hou <ha...@apache.org> wrote:
> >
> > > Hi,
> > >
> > > I have a draft of 2.0.10 release notes. Let me list here.
> > > Features & Improvements
> > > - Data Query: Support schema-level and table-level storage space
> statistics
> > > - Data Query: Add DEBUG SQL capability and optimize the result set of
> > > Explain Analyze
> > > - Data Query: Support set operations (UNION/INTERSECT/EXCEPT) and
> Common
> > > Table Expressions (CTE) in the table model
> > > - Data Query: Add IF scalar function, binary functions and
> > > APPROX_PERCENTILE aggregate function for the table model
> > > - Data Query: Optimize display fields of the show queries command, add
> > > three new columns: client IP that initiates the query, timeout
> threshold,
> > > and total request waiting latency
> > > - Storage Management: Provide C-language driver SDK interfaces,
> including
> > > parameter binding, automatic failover for multi-node switching and
> other
> > > fault-tolerant capabilities
> > > - System Management: Add the capability to manually trigger node load
> > > balancing
> > > - System Management: Add progress query capability for the remove
> datanode
> > > operation
> > > - System Management: Introduce the show configuration statement to view
> > > cluster configuration information
> > > - System Management: Add statistics for the number of slow write
> requests
> > > - AI Management: Add two built-in models: Moirai2 and Toto
> > > - AI Management: Enable AINode to manage multiple models under the same
> > > model_type
> > > - AI Management: Support loading built-in models for CPU inference with
> > > isolated inference processes
> > >
> > > Bugs
> > > - Fix the issue that partition tables get lost when leader switch
> happens
> > > on single-node ConfigNode
> > > - Fix the bug where configurations delivered via set configuration to a
> > > single DataNode take effect on all other DataNodes unexpectedly
> > > - Fix the occasional continuous accumulation of reservedSize for
> > > timeout-only queries across multiple devices in tree model, which
> triggers
> > > query errors after hitting the max capacity limit
> > > - Fix the parsing error in tree-model-to-table conversion logic when
> tree
> > > paths consist entirely of numeric characters
> > > - Correct abnormal output of show devices root.db.** when root.db
> itself
> > > is a device, making it display root.db properly
> > > - Fix the failure of kill query and query timeout mechanism when the
> > > client holds the connection open without fully consuming all result
> sets
> > > - Fix array index out-of-bounds error caused by null values in
> attribute
> > > fields on the receiver side, which blocks the data synchronization
> pipeline
> > > - Fix connection pool leakage exceptions in the C# client
> > > - Fix exceptions occurring in the C++ client when write redirection is
> > > enabled while dn_rpc_address is configured to 0.0.0.0 on DataNode
> server
> > > - Fix abnormal repeated broadcasting of deletion markers (Deletions)
> > > during the load operation
> > > - Fix the issue that rows with null inserted fields cannot be
> synchronized
> > > to the receiver in real-time data synchronization scenarios
> > > - Fix incorrect idempotency logic for out-of-order TTL-expired data on
> the
> > > synchronization receiver, which leads to repeated synchronization of
> > > expired data
> > > - Fix the problem that enable-send-tsfile-limit parameter cannot be
> > > manually configured for historical Pipes split from full-data Pipe
> tasks
> > > - Fix Pipe temporary stoppage triggered by insufficient memory of a
> single
> > > subtask within synchronization jobs
> > > - Fix the issue that modifying username/password of source or sink in
> > > write-back-sink model generates error logs and prevents data from being
> > > received on the sink side
> > > - Fix permission issue preventing non-root users from stopping the
> AINode
> > > process
> > > - Mitigate memory leaks that tend to occur during long-running
> continuous
> > > inference of models running on CPU
> > >
> > > BR,
> > > Haonan
> > >
> > >
> > > On 2026/06/30 09:38:46 Yuan Tian wrote:
> > > > Hi all,
> > > >
> > > > We would like to skip the IoTDB 2.0.9 release and proceed directly
> with
> > > the
> > > > 2.0.10 release.
> > > >
> > > > The main reason is that one release-note item originally planned for
> > > 2.0.9
> > > > remained under active development and iteration longer than
> expected. By
> > > > the time this item was fully completed, the development work planned
> for
> > > > 2.0.10 had also been finished.
> > > >
> > > > Given this situation, releasing 2.0.9 separately would bring limited
> > > > practical value and may instead cause unnecessary confusion.
> Therefore,
> > > we
> > > > plan to include the relevant changes in 2.0.10 and make 2.0.10 the
> next
> > > > release version.
> > > >
> > > > To make future release planning clearer, starting from 2.0.11, before
> > > each
> > > > release iteration begins, we will send an email to the mailing list
> > > > outlining the release-note items expected to be included in that
> version,
> > > > together with the expected code-freeze date. Everyone is also
> welcome to
> > > > propose additional release-note items that should be included in the
> > > > corresponding version. Please note that the code-freeze date is not
> the
> > > > same as the release date. After code freeze, we will still need to
> test
> > > and
> > > > verify the corresponding release-note items before proceeding with
> the
> > > > formal release.
> > > >
> > > > Please let us know if you have any concerns or objections.
> > > >
> > > > Best regards,
> > > >
> > > > ------------------------
> > > >
> > > > Yuan Tian
> > > >
> > >
> >
>
Re: Proposal to Skip IoTDB 2.0.9 and Release 2.0.10 Directly
Posted by Haonan Hou <ha...@apache.org>.
Hi Yuan,
I have submit a PR [1] to do it. Please take a look.
[1] https://github.com/apache/iotdb/pull/18093
Best,
Haonan
On 2026/07/02 08:53:53 Yuan Tian wrote:
> Hi haonan,
>
> LGTM, since rc/2.0.10 has already been checked out, I think we can bump
> the version of master to 2.0.11-SNAPSHOT?
>
>
> Best regards,
> ---------------------
> Yuan Tian
>
> On Thu, Jul 2, 2026 at 4:47 PM Haonan Hou <ha...@apache.org> wrote:
>
> > Hi,
> >
> > I have a draft of 2.0.10 release notes. Let me list here.
> > Features & Improvements
> > - Data Query: Support schema-level and table-level storage space statistics
> > - Data Query: Add DEBUG SQL capability and optimize the result set of
> > Explain Analyze
> > - Data Query: Support set operations (UNION/INTERSECT/EXCEPT) and Common
> > Table Expressions (CTE) in the table model
> > - Data Query: Add IF scalar function, binary functions and
> > APPROX_PERCENTILE aggregate function for the table model
> > - Data Query: Optimize display fields of the show queries command, add
> > three new columns: client IP that initiates the query, timeout threshold,
> > and total request waiting latency
> > - Storage Management: Provide C-language driver SDK interfaces, including
> > parameter binding, automatic failover for multi-node switching and other
> > fault-tolerant capabilities
> > - System Management: Add the capability to manually trigger node load
> > balancing
> > - System Management: Add progress query capability for the remove datanode
> > operation
> > - System Management: Introduce the show configuration statement to view
> > cluster configuration information
> > - System Management: Add statistics for the number of slow write requests
> > - AI Management: Add two built-in models: Moirai2 and Toto
> > - AI Management: Enable AINode to manage multiple models under the same
> > model_type
> > - AI Management: Support loading built-in models for CPU inference with
> > isolated inference processes
> >
> > Bugs
> > - Fix the issue that partition tables get lost when leader switch happens
> > on single-node ConfigNode
> > - Fix the bug where configurations delivered via set configuration to a
> > single DataNode take effect on all other DataNodes unexpectedly
> > - Fix the occasional continuous accumulation of reservedSize for
> > timeout-only queries across multiple devices in tree model, which triggers
> > query errors after hitting the max capacity limit
> > - Fix the parsing error in tree-model-to-table conversion logic when tree
> > paths consist entirely of numeric characters
> > - Correct abnormal output of show devices root.db.** when root.db itself
> > is a device, making it display root.db properly
> > - Fix the failure of kill query and query timeout mechanism when the
> > client holds the connection open without fully consuming all result sets
> > - Fix array index out-of-bounds error caused by null values in attribute
> > fields on the receiver side, which blocks the data synchronization pipeline
> > - Fix connection pool leakage exceptions in the C# client
> > - Fix exceptions occurring in the C++ client when write redirection is
> > enabled while dn_rpc_address is configured to 0.0.0.0 on DataNode server
> > - Fix abnormal repeated broadcasting of deletion markers (Deletions)
> > during the load operation
> > - Fix the issue that rows with null inserted fields cannot be synchronized
> > to the receiver in real-time data synchronization scenarios
> > - Fix incorrect idempotency logic for out-of-order TTL-expired data on the
> > synchronization receiver, which leads to repeated synchronization of
> > expired data
> > - Fix the problem that enable-send-tsfile-limit parameter cannot be
> > manually configured for historical Pipes split from full-data Pipe tasks
> > - Fix Pipe temporary stoppage triggered by insufficient memory of a single
> > subtask within synchronization jobs
> > - Fix the issue that modifying username/password of source or sink in
> > write-back-sink model generates error logs and prevents data from being
> > received on the sink side
> > - Fix permission issue preventing non-root users from stopping the AINode
> > process
> > - Mitigate memory leaks that tend to occur during long-running continuous
> > inference of models running on CPU
> >
> > BR,
> > Haonan
> >
> >
> > On 2026/06/30 09:38:46 Yuan Tian wrote:
> > > Hi all,
> > >
> > > We would like to skip the IoTDB 2.0.9 release and proceed directly with
> > the
> > > 2.0.10 release.
> > >
> > > The main reason is that one release-note item originally planned for
> > 2.0.9
> > > remained under active development and iteration longer than expected. By
> > > the time this item was fully completed, the development work planned for
> > > 2.0.10 had also been finished.
> > >
> > > Given this situation, releasing 2.0.9 separately would bring limited
> > > practical value and may instead cause unnecessary confusion. Therefore,
> > we
> > > plan to include the relevant changes in 2.0.10 and make 2.0.10 the next
> > > release version.
> > >
> > > To make future release planning clearer, starting from 2.0.11, before
> > each
> > > release iteration begins, we will send an email to the mailing list
> > > outlining the release-note items expected to be included in that version,
> > > together with the expected code-freeze date. Everyone is also welcome to
> > > propose additional release-note items that should be included in the
> > > corresponding version. Please note that the code-freeze date is not the
> > > same as the release date. After code freeze, we will still need to test
> > and
> > > verify the corresponding release-note items before proceeding with the
> > > formal release.
> > >
> > > Please let us know if you have any concerns or objections.
> > >
> > > Best regards,
> > >
> > > ------------------------
> > >
> > > Yuan Tian
> > >
> >
>
Re: Proposal to Skip IoTDB 2.0.9 and Release 2.0.10 Directly
Posted by Yuan Tian <ja...@gmail.com>.
Hi haonan,
LGTM, since rc/2.0.10 has already been checked out, I think we can bump
the version of master to 2.0.11-SNAPSHOT?
Best regards,
---------------------
Yuan Tian
On Thu, Jul 2, 2026 at 4:47 PM Haonan Hou <ha...@apache.org> wrote:
> Hi,
>
> I have a draft of 2.0.10 release notes. Let me list here.
> Features & Improvements
> - Data Query: Support schema-level and table-level storage space statistics
> - Data Query: Add DEBUG SQL capability and optimize the result set of
> Explain Analyze
> - Data Query: Support set operations (UNION/INTERSECT/EXCEPT) and Common
> Table Expressions (CTE) in the table model
> - Data Query: Add IF scalar function, binary functions and
> APPROX_PERCENTILE aggregate function for the table model
> - Data Query: Optimize display fields of the show queries command, add
> three new columns: client IP that initiates the query, timeout threshold,
> and total request waiting latency
> - Storage Management: Provide C-language driver SDK interfaces, including
> parameter binding, automatic failover for multi-node switching and other
> fault-tolerant capabilities
> - System Management: Add the capability to manually trigger node load
> balancing
> - System Management: Add progress query capability for the remove datanode
> operation
> - System Management: Introduce the show configuration statement to view
> cluster configuration information
> - System Management: Add statistics for the number of slow write requests
> - AI Management: Add two built-in models: Moirai2 and Toto
> - AI Management: Enable AINode to manage multiple models under the same
> model_type
> - AI Management: Support loading built-in models for CPU inference with
> isolated inference processes
>
> Bugs
> - Fix the issue that partition tables get lost when leader switch happens
> on single-node ConfigNode
> - Fix the bug where configurations delivered via set configuration to a
> single DataNode take effect on all other DataNodes unexpectedly
> - Fix the occasional continuous accumulation of reservedSize for
> timeout-only queries across multiple devices in tree model, which triggers
> query errors after hitting the max capacity limit
> - Fix the parsing error in tree-model-to-table conversion logic when tree
> paths consist entirely of numeric characters
> - Correct abnormal output of show devices root.db.** when root.db itself
> is a device, making it display root.db properly
> - Fix the failure of kill query and query timeout mechanism when the
> client holds the connection open without fully consuming all result sets
> - Fix array index out-of-bounds error caused by null values in attribute
> fields on the receiver side, which blocks the data synchronization pipeline
> - Fix connection pool leakage exceptions in the C# client
> - Fix exceptions occurring in the C++ client when write redirection is
> enabled while dn_rpc_address is configured to 0.0.0.0 on DataNode server
> - Fix abnormal repeated broadcasting of deletion markers (Deletions)
> during the load operation
> - Fix the issue that rows with null inserted fields cannot be synchronized
> to the receiver in real-time data synchronization scenarios
> - Fix incorrect idempotency logic for out-of-order TTL-expired data on the
> synchronization receiver, which leads to repeated synchronization of
> expired data
> - Fix the problem that enable-send-tsfile-limit parameter cannot be
> manually configured for historical Pipes split from full-data Pipe tasks
> - Fix Pipe temporary stoppage triggered by insufficient memory of a single
> subtask within synchronization jobs
> - Fix the issue that modifying username/password of source or sink in
> write-back-sink model generates error logs and prevents data from being
> received on the sink side
> - Fix permission issue preventing non-root users from stopping the AINode
> process
> - Mitigate memory leaks that tend to occur during long-running continuous
> inference of models running on CPU
>
> BR,
> Haonan
>
>
> On 2026/06/30 09:38:46 Yuan Tian wrote:
> > Hi all,
> >
> > We would like to skip the IoTDB 2.0.9 release and proceed directly with
> the
> > 2.0.10 release.
> >
> > The main reason is that one release-note item originally planned for
> 2.0.9
> > remained under active development and iteration longer than expected. By
> > the time this item was fully completed, the development work planned for
> > 2.0.10 had also been finished.
> >
> > Given this situation, releasing 2.0.9 separately would bring limited
> > practical value and may instead cause unnecessary confusion. Therefore,
> we
> > plan to include the relevant changes in 2.0.10 and make 2.0.10 the next
> > release version.
> >
> > To make future release planning clearer, starting from 2.0.11, before
> each
> > release iteration begins, we will send an email to the mailing list
> > outlining the release-note items expected to be included in that version,
> > together with the expected code-freeze date. Everyone is also welcome to
> > propose additional release-note items that should be included in the
> > corresponding version. Please note that the code-freeze date is not the
> > same as the release date. After code freeze, we will still need to test
> and
> > verify the corresponding release-note items before proceeding with the
> > formal release.
> >
> > Please let us know if you have any concerns or objections.
> >
> > Best regards,
> >
> > ------------------------
> >
> > Yuan Tian
> >
>