Skip to content

Releases: taosdata/TDengine

3.0.5.0

08 Jun 01:52
Compare
Choose a tag to compare

Dear community members,

The past six years have been a long but exciting journey for TDengine, seeing the evolution of our product from a basic TSDB to a full-featured cloud-native solution for time-series data, and the growth of our team from a five-person startup to the 80 members of today.
Throughout our journey, our constant goal has been to provide our community with first-class products and services. Today, on TDengine’s sixth anniversary, it is with much pride that we look back on the evolution of TDengine from 1.0 to 2.0 and 3.0, from the first line of code written by Jeff himself to the hundreds of thousands of lines that now comprise our product, indicating the continuous growth of our ecosystem, the increasing maturity of features such as stream processing and data subscription, the launch of our fully managed cloud service – and, of course, the resolution of many issues brought up by community members like yourself.
And indeed, what gives our team the most satisfaction is the growth of our open-source community over the years, exemplified by our over 21,000 stars on GitHub and more than 250,000 instances installed worldwide. We are delighted that an ever-increasing number of users and developers have chosen to become members of our community and accompany us on our journey, and we would like to take this opportunity to express our sincere gratitude for your support – our greatest motivation to continue our work and accomplish our goals.
Going forward, we intend to continue innovating in the time-series database field and enabling digital transformation across more industries. It is only through your support that we will be able to deepen our understanding of the scenarios and use cases of various industries and to maintain the growth of our product in a way that aligns with the needs of our community. Once again, thank you for being there with us over the past six years, and we hope that we can count on your continuing support in the future.

The TDengine Team, 2023/Jun/6

New Features and Improvements in TDengine 3.0.5.0

  1. System stability & performance
  • Improved system stability under high stress data writing
  • Optimized system performance in some query scenarios
  • Altering database replicas doesn't block writing by introducing RAFT Learner
  • Write driven cache for last() and last_row() to improve the query performance
  • Optimized time cost of creating/dropping database
  • Log long queries by default for easy debugging
  • Controlled meta data cache in taosc library
  • dnode can be restored after its data is totally lost (Enterprise only)
  1. System security
  • Privilege control at table level (Enterprise only)
  • License key can be updated using SQL command by "root" (Enterprise only)
  1. Stream processing
  • Significantly reduced disk I/O and memory usage
  • Stream can be paused/resumed
  1. Data Subscription
  • Consuming progress can be queried
  • Consumers can perform seek operation
  • Consumers can subscribe supertable with tag filtering
  • Consumers can retrieve meta data based on a topic name
  • Improved performance
  1. Others
  • Maximum row length is increased to 64KB
  • interp() can be used for super table
  • Python UDF can support multiple versions with "REPLACE" command
  • Partition by and window clause can be followed by "Having" clause

What's Changed

Read more

3.0.4.2

25 May 03:29
Compare
Choose a tag to compare

Enhancements:

  1. Improved performance for last() without cache enabled
  2. Minimize the impact on data writing caused by data subscription
  3. Improved rebalance performance

Fixed bugs:

  1. taosd exit may take too much time when there is a long query
  2. data subscription may hang in some extreme cases
  3. order by _wstart with partition by doesn't work
  4. A few crashe problems when using stream
  5. The result of sum + case when is not correct
  6. The result of interp is incorrect
  7. Schemaless writing reported error "table not exist"
  8. taosd keeps logging "database not exist"

Full Changelog: ver-3.0.4.2...ver-3.0.4.2

What's Changed

Read more

3.0.4.1

28 Apr 01:03
Compare
Choose a tag to compare

Enhancements

  • Performance improvement for data subscription
  • User experience improvement of quitting taos shell
  • Topic name of mixed lower and upper case can be subscribed
  • Output format of float/double are synchronized

Fixed bugs

  • "alter replica 3" hangs in some conditions
  • "trim/flush database" doesn't work with multi-level storage
  • Windows client can generate log file bigger than 4GB
  • Some crash issues

What's Changed

New Contributors

Full Changelog: ver-3.0.4.0...ver-3.0.4.1

3.0.4.0

17 Apr 02:18
b3acd9b
Compare
Choose a tag to compare

Release Notes

  1. Significantly improved stability
  2. Improved performance in some query scenarios: 1) the performance of mode() when used with interval(); 2) the performance of percentile(); 3) the performance of last()/last_row()
  3. Rebalancing ability of cluster workloads after restarting one dnode
  4. Improved robustness when the data of a dnode is disrupted
  5. Data compact can be performed based on a specified time range
  6. Database parameters stt_trigger and minRows can be changed dynamically
  7. User defined function can be developed in Python
  8. Data in Pi, OPC UA and OPC DA can be transferred into TDengine
  9. taosExplorer can help manage data sources and tasks for Pi, OPC UA and OPC DA

What's Changed

Read more

3.0.3.2

06 Apr 09:13
Compare
Choose a tag to compare

Enhancements:

  • Improved stability, especially in case of high system load
  • Performance improvement when querying "last() group by tbname"
  • Memory usage optimization of data subscription
  • Perforamnce improvement for last() and last_row()
  • Optimizaed WAL cleanup for mnode
  • Performance improvement for spread()

Fixed bugs:

  • taosd error caused by data transfering using taosX
  • dnode offline during the process of alter replica
  • Client hangs after consuming all messages and then restart
  • The result of FILL(LINEAR) is not consistent for incoming data and existing data
  • Max() returns incorrect result
  • Some other bugs

What's Changed

Full Changelog: ver-3.0.3.1...ver-3.0.3.2

3.0.3.1

22 Mar 02:45
Compare
Choose a tag to compare

Enhancements:

  • Adjusted the behavior of limit/slimit with group by to be compatible with 2.x
  • Performance improvement for count(1)

Fixed bugs:

  • non-root user query information_schema.ins_dnodes caused client crash
  • group by tag with limit doesn't work
  • last() result is not correct when queryPolicy is set to 3
  • filter for non-key timestamp caused too many window
  • Errors reported in writing with schemaless
  • Stream doens't work after the table schema was altered
  • Group by doesn't return expected result
  • Improper error message was reported when using order by

What's Changed

Read more

3.0.3.0

10 Mar 05:47
4c69835
Compare
Choose a tag to compare

New Features:

  • You can now create windowed queries that are triggered by events - just specify any conditions that TDengine supports, including operations on different columns.
  • Indexing on any tag: You can now create an index on any tag, enabling improved query performance when filtering by tag.
  • (Enterprise Edition only) Data compaction: TDengine Enterprise Edition now provides the COMPACT statement to reorganize your database and remove gaps from deleted or updated records. This can greatly improve storage and query performance on deployments that have been running for some time.
  • (Enterprise Edition only) TDengine Explorer: TDengine Enterprise Edition now offers a convenient GUI to make ad hoc time-series operations and instance management even easier.
  • (Enterprise Edition only) Local backup and migration: The latest version of taosX included with TDengine Enterprise Edition 3.0.3.0 can back up and restore a TDengine deployment from a local file. taosX can also migrate data from existing TDengine 2.6 or 3.x deployments to TDengine 3.0.3.0, either incrementally or in full.

Enhancements:

  • Results of stream processing can be stored in existing stables
  • You can specify tags when storing the results of stream processing
  • Data can be subscribed over websocket using Java Connector
  • Data can be subscribed over websocket using Python Connector
  • You can monitor HTTP status code on Grafana Dashboard
  • Grafana plugin supports multi-dimensional alerts
  • taosBenchmark can write data into sepcified range of child tables
  • taosBenchmark can specify the number of vgroups on command line

What's Changed

Read more

3.0.2.6

01 Mar 08:56
Compare
Choose a tag to compare

Enhancements:
N/A

Fixed Bugs:

  • The result of max() are not consistent between first + max + interval(1d) and last + max + interval(1d)
  • join across multiple databases hangs
  • invalid values of consumer_id and vgroup_id are returned in "show subscriptions"
  • data ingestion may be blocked because of many concurrent queries
  • taosd OOM after long time running with stream and data ingestion
  • stable can't be deleted even there isn't topic on it
  • "table doesn't exist" is returned when "show table distributed" for an existing normal table
  • Result is wrong when querying "order by ts desc"
  • Result is wrong when querying last_row() on a supertable

3.0.2.5

10 Feb 08:16
Compare
Choose a tag to compare

Enhancements:

  • tar.gz installer can detect RPM/Deb installation and remove them per user's confirmation
  • Performance improvement of partition by
  • Performance improvement of "select * limit 0"
  • Performance improvement of "select count(*)" when there is out-of-order data
  • Performance improvement for query by caching head files

Fixed Bugs:

  • taosd crashs when querying information_schema with filter conditions
  • taosd crashs when inserting data rows with duplicate timestamp in some extreme cases
  • taosd crashs when executing some very complicated queries
  • taosd crashs on Windows
  • keepColumnName doesn't work in sub-query
  • min(), max() and some functions don't return correct result when all data are in data files
  • taoskeeper failed to start because "create database" is not executed prior to "create table" in some cases
  • Some memory leak problems in taosd
  • Memory leak in taosadpater when using line protocol
  • No result is returned using "order by asc limit offset"
  • interp() doesn't work properly in sub-query
  • Error "WAL failed to remove log file" on Windows
  • A lot of internal bugs to improve stability

3.0.2.4

29 Jan 08:38
Compare
Choose a tag to compare

Improvements:

  • "show cluster" can display "version" and "expired" fields

Fixed Bugs:

  • "show create database" doesn't display "STT_TRIGGER" parameter
  • Application using JAVA connector crashed after upgrading to 3.0.2.3
  • "insert into d1 select * from d0" caused taosd hang
  • oom of taosadapter while running data replication