Re: [jira] [Created] (IOTDB-229) Inconsistent usage of Marker in TsFile
+1, generally the javadoc is not incorrect but can be written more clearly. Lei Rui On 9/19/2019 14:36,Xiangdong Huang wrote: Hi, Ok then just take care when using it. No, I think we need to fix the javadoc. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Lei Rui 于2019年9月19日周四 下午2:31写道: Hi, I do not know why the mail does not show the description... Sorry it is because I didn't know how to add the detail description when opening an issue on JIRA. Thankfully, I now know how to do that. Below is the raw description: ``` 1) Structure ChunkMetaData.offsetOfChunkHeader refers to the position of Marker_1, while TsDeviceMetadataIndex refers to the position AFTER Marker_2. This may impose unnecessary learning burdens on users. As suggested by Jialin Qiao, an ideal case may be that ChunkMetaData.offsetOfChunkHeader refers to the position AFTER Marker_1 too. 2) Annotation The parameter `boolean markerRead` is used across multiple methods. I found two annotations of it. One annotates "read marker (boolean type)" and the other annotates "Whether the marker of the CHUNK_HEADER has been read". When `markerRead` is FALSE, the former usage will add the current read offset by 1 while the latter usage will read the byte of the Marker. Although they both push forward the read process, the two different meanings of `markerRead` is a bit of a counter-intuitive. A solution is changing the name of the parameter from `markerRead` to `isMarkerExist`, and then when `isMarkerExist` is TRUE, the former usage can choose to add the current read offset by 1 to skip it while the latter usage can choose to read the byte of the Marker. Remind to maintain the tsfile/format-changelist.md if changes are made. ``` If it is true, then the reader do not need to read 1 byte. Ok then just take care when using it. Lei Rui On 9/19/2019 14:12,Xiangdong Huang wrote: Hi, I do not know why the mail does not show the description... As suggested by Jialin Qiao, an ideal case may be that ChunkMetaData.offsetOfChunkHeader refers to the position AFTER Marker_1 too. +1. A solution is changing the name of the parameter from `markerRead` to `isMarkerExist` The correct meaning of the parameter should be `has the Marker been read`. If it is true, then the reader do not need to read 1 byte. I suggest to make the inconsistency to the above meaning. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Lei Rui (Jira) 于2019年9月19日周四 下午1:27写道: Lei Rui created IOTDB-229: - Summary: Inconsistent usage of Marker in TsFile Key: IOTDB-229 URL: https://issues.apache.org/jira/browse/IOTDB-229 Project: Apache IoTDB Issue Type: Improvement Reporter: Lei Rui -- This message was sent by Atlassian Jira (v8.3.4#803005)
Re: [jira] [Created] (IOTDB-229) Inconsistent usage of Marker in TsFile
Hi, > Ok then just take care when using it. No, I think we need to fix the javadoc. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Lei Rui 于2019年9月19日周四 下午2:31写道: > Hi, > > > > I do not know why the mail does not show the description... > Sorry it is because I didn't know how to add the detail description when > opening an issue on JIRA. Thankfully, I now know how to do that. > > > Below is the raw description: > ``` > 1) Structure > ChunkMetaData.offsetOfChunkHeader refers to the position of Marker_1, > while TsDeviceMetadataIndex refers to the position AFTER Marker_2. This may > impose unnecessary learning burdens on users. > As suggested by Jialin Qiao, an ideal case may be that > ChunkMetaData.offsetOfChunkHeader refers to the position AFTER Marker_1 too. > 2) Annotation > The parameter `boolean markerRead` is used across multiple methods. I > found two annotations of it. One annotates "read marker (boolean type)" and > the other annotates "Whether the marker of the CHUNK_HEADER has been read". > When `markerRead` is FALSE, the former usage will add the current read > offset by 1 while the latter usage will read the byte of the Marker. > Although they both push forward the read process, the two different > meanings of `markerRead` is a bit of a counter-intuitive. > A solution is changing the name of the parameter from `markerRead` to > `isMarkerExist`, and then when `isMarkerExist` is TRUE, the former usage > can choose to add the current read offset by 1 to skip it while the latter > usage can choose to read the byte of the Marker. > Remind to maintain the tsfile/format-changelist.md if changes are made. > ``` > > > > > > If it is true, then the reader do not need to read 1 byte. > Ok then just take care when using it. > > > Lei Rui > > > > > > > On 9/19/2019 14:12,Xiangdong Huang wrote: > Hi, > > I do not know why the mail does not show the description... > > As suggested by Jialin Qiao, an ideal case may be that > ChunkMetaData.offsetOfChunkHeader refers to the position AFTER Marker_1 > too. > > +1. > > A solution is changing the name of the parameter from `markerRead` to > `isMarkerExist` > > The correct meaning of the parameter should be `has the Marker been read`. > If it is true, then the reader do not need to read 1 byte. > > I suggest to make the inconsistency to the above meaning. > > Best, > --- > Xiangdong Huang > School of Software, Tsinghua University > > 黄向东 > 清华大学 软件学院 > > > Lei Rui (Jira) 于2019年9月19日周四 下午1:27写道: > > Lei Rui created IOTDB-229: > - > > Summary: Inconsistent usage of Marker in TsFile > Key: IOTDB-229 > URL: https://issues.apache.org/jira/browse/IOTDB-229 > Project: Apache IoTDB > Issue Type: Improvement > Reporter: Lei Rui > > > > > > > -- > This message was sent by Atlassian Jira > (v8.3.4#803005) > >
Re: [jira] [Created] (IOTDB-229) Inconsistent usage of Marker in TsFile
Hi, > I do not know why the mail does not show the description... Sorry it is because I didn't know how to add the detail description when opening an issue on JIRA. Thankfully, I now know how to do that. Below is the raw description: ``` 1) Structure ChunkMetaData.offsetOfChunkHeader refers to the position of Marker_1, while TsDeviceMetadataIndex refers to the position AFTER Marker_2. This may impose unnecessary learning burdens on users. As suggested by Jialin Qiao, an ideal case may be that ChunkMetaData.offsetOfChunkHeader refers to the position AFTER Marker_1 too. 2) Annotation The parameter `boolean markerRead` is used across multiple methods. I found two annotations of it. One annotates "read marker (boolean type)" and the other annotates "Whether the marker of the CHUNK_HEADER has been read". When `markerRead` is FALSE, the former usage will add the current read offset by 1 while the latter usage will read the byte of the Marker. Although they both push forward the read process, the two different meanings of `markerRead` is a bit of a counter-intuitive. A solution is changing the name of the parameter from `markerRead` to `isMarkerExist`, and then when `isMarkerExist` is TRUE, the former usage can choose to add the current read offset by 1 to skip it while the latter usage can choose to read the byte of the Marker. Remind to maintain the tsfile/format-changelist.md if changes are made. ``` > If it is true, then the reader do not need to read 1 byte. Ok then just take care when using it. Lei Rui On 9/19/2019 14:12,Xiangdong Huang wrote: Hi, I do not know why the mail does not show the description... As suggested by Jialin Qiao, an ideal case may be that ChunkMetaData.offsetOfChunkHeader refers to the position AFTER Marker_1 too. +1. A solution is changing the name of the parameter from `markerRead` to `isMarkerExist` The correct meaning of the parameter should be `has the Marker been read`. If it is true, then the reader do not need to read 1 byte. I suggest to make the inconsistency to the above meaning. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Lei Rui (Jira) 于2019年9月19日周四 下午1:27写道: Lei Rui created IOTDB-229: - Summary: Inconsistent usage of Marker in TsFile Key: IOTDB-229 URL: https://issues.apache.org/jira/browse/IOTDB-229 Project: Apache IoTDB Issue Type: Improvement Reporter: Lei Rui -- This message was sent by Atlassian Jira (v8.3.4#803005)
Re: Resolve "input line is too long" in windows os
Hi, JVM loads a set of jars according to the order of users declaration. If you set classpath=log4j-1.1.jar : log4j-0.9.jar : log4j-1.2.jar, then JVM will load log4-1.1.jar first. If there are duplicate files among the 3 jars, then only the first one will be loaded into JVM. That is why we must set it as classpath= %iotdb classpath% : %system classpath%. Though for current version of IoTDB, you can just set classpath= %iotdb classpath%, I'd like to do not cover system classpath, because it is more general. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 JT 于2019年9月19日周四 下午2:01写道: > Hi, > > Should we really include the original CLASSPATH? We may get ourselves into > trouble if it contains some incompatible or conflict dependencies with > those in IoTDB. For example, if the user has put an old Log4j into > CLASSPATH, the system log may fail. > > Tian Jiang > > 发送自 Windows 10 版邮件应用 > > 发件人: Lei Rui > 发送时间: 2019年9月19日 13:42 > 收件人: dev@iotdb.apache.org > 主题: Re: Resolve "input line is too long" in windows os > > Hi, > > > Ok then it looks like `start-server.bat` needs changing because now it is > > > ``` > @REM * CLASSPATH library setting * > @REM Ensure that any user defined CLASSPATH variables are not used on > startup > set CLASSPATH="%IOTDB_HOME%\lib" > @REM For each jar in the IOTDB_HOME lib directory call append to build the > CLASSPATH variable. > set CLASSPATH=%CLASSPATH%;"%IOTDB_HOME%\lib\*" > set CLASSPATH=%CLASSPATH%;iotdb.IoTDB > goto okClasspath > ``` > > > Lei Rui > On 9/19/2019 13:33,Xiangdong Huang wrote: > Hi, > > set CLASSPATH=%CLASSPATH%;%IOTDB_CLI_HOME%\lib\* > > Normally, the better way is : > > set CLASSPATH=%IOTDB_CLI_HOME%\lib\*;%CLASSPATH% > > It means load %IOTDB_CLI_HOME%\lib\* first, and then %CLASSPATH% that the > system original having. > > Best, > --- > Xiangdong Huang > School of Software, Tsinghua University > > 黄向东 > 清华大学 软件学院 > > > Jialin Qiao 于2019年9月5日周四 上午9:21写道: > > > Hanot iotdb-env.sh, it is in the start-client.bat, import-csv.bat and > export-csv.bat... > > Best > -- > Jialin Qiao > School of Software, Tsinghua University > > 乔嘉林 > 清华大学 软件学院 > > -原始邮件- > 发件人: "Jialin Qiao" > 发送时间: 2019-09-05 09:17:45 (星期四) > 收件人: dev@iotdb.apache.org > 抄送: > 主题: Re: Resolve "input line is too long" in windows os > > Hi, > > Ok, I found this in the current iotdb-env.sh: > > ``` > @REM Ensure that any user defined CLASSPATH variables are not used on > startup > set CLASSPATH="" > ``` > > Best, > -- > Jialin Qiao > School of Software, Tsinghua University > > 乔嘉林 > 清华大学 软件学院 > > -原始邮件- > 发件人: "Tian Jiang" > 发送时间: 2019-09-05 09:05:29 (星期四) > 收件人: "dev@iotdb.apache.org" > 抄送: "dev@iotdb.apache.org" > 主题: Re: Resolve "input line is too long" in windows os > > Hi, > > > I think it is fine, since we are not sure what %CLASSPATH% would be > originally, if we depend on some unknown external dependencies, things may > get uncontrollable. > > > Tian Jiang > > > | | > Tian Jiang > | > | > jt2594...@163.com > | > 签名由网易邮箱大师定制 > On 9/5/2019 08:55,Jialin Qiao wrote: > Hi Zesong, > > Thanks for sharing your experience :) > > I just have a question about this line: > > set CLASSPATH=%IOTDB_CLI_HOME%\lib\* > > Should it be the following command? > > set CLASSPATH=%CLASSPATH%;%IOTDB_CLI_HOME%\lib\* > > Best, > -- > Jialin Qiao > School of Software, Tsinghua University > > 乔嘉林 > 清华大学 软件学院 > > -原始邮件- > 发件人: "Zesong Sun" > 发送时间: 2019-09-04 22:20:54 (星期三) > 收件人: dev > 抄送: > 主题: Resolve "input line is too long" in windows os > > Hi, > > > > > Recently I have encountered a problem when updating maven dependency > in tsfile module: when starting client in windows os, I will get an error: > "The input line is too long." > > > It is resulted by limitation of the maximum length of the string that > can be used at the command line: "On computers running Microsoft Windows XP > or later, the maximum length of the string that you can use at the command > prompt is 8191 characters. On computers running Microsoft Windows 2000 or > Windows NT 4.0, the maximum length of the string that you can use at the > command prompt is 2047 characters. This limitation applies to the command > line, individual environment variables (such as the PATH variable) that are > inherited by other processes, and all environment variable expansions. If > you use Command Prompt to run batch files, this limitation also applies to > batch file processing.” [1] > > > With the suggestion of @Jialin Qiao, I reduced the dependency (from > hadoop-client to hadoop-common), and local tests are all passed. However, > the tests on Travis CI still failed. > > > Thanks to @Tian Jiang, I changed the set CLASSPATH statement in .bat > file of starting client: > > > Previous: > ``` > for %%i in ("%IOTDB_CLI_HOME%\lib\*.jar") do call :append "%%i" > goto okClasspath > > > :append > set CLASSPATH=%CLASSPATH%;%1 > goto
Re: [jira] [Created] (IOTDB-223) TsFile skeleton viewer
Hi, Yes, adding a script is more user-friendly. I'll find time to do that. By the way, do you guys have any opinions on this tool names, which one is better? "TsFile skeleton viewer" or "TsFile sketch tool"? Lei Rui On 9/19/2019 13:48,Xiangdong Huang wrote: Hi, I like the idea, and I notice the PR has been merged. A suggestion: as you consider the skeleton viewer as a tool. how about moving the java file into "tool" package, and add a script to use it? Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Lei Rui (Jira) 于2019年9月19日周四 上午12:42写道: Lei Rui created IOTDB-223: - Summary: TsFile skeleton viewer Key: IOTDB-223 URL: https://issues.apache.org/jira/browse/IOTDB-223 Project: Apache IoTDB Issue Type: Improvement Reporter: Lei Rui -- This message was sent by Atlassian Jira (v8.3.4#803005)
Re: [jira] [Created] (IOTDB-229) Inconsistent usage of Marker in TsFile
Hi, I do not know why the mail does not show the description... > As suggested by Jialin Qiao, an ideal case may be that ChunkMetaData.offsetOfChunkHeader refers to the position AFTER Marker_1 too. +1. > A solution is changing the name of the parameter from `markerRead` to `isMarkerExist` The correct meaning of the parameter should be `has the Marker been read`. If it is true, then the reader do not need to read 1 byte. I suggest to make the inconsistency to the above meaning. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Lei Rui (Jira) 于2019年9月19日周四 下午1:27写道: > Lei Rui created IOTDB-229: > - > > Summary: Inconsistent usage of Marker in TsFile > Key: IOTDB-229 > URL: https://issues.apache.org/jira/browse/IOTDB-229 > Project: Apache IoTDB > Issue Type: Improvement > Reporter: Lei Rui > > > > > > > -- > This message was sent by Atlassian Jira > (v8.3.4#803005) >
答复: Resolve "input line is too long" in windows os
Hi, Should we really include the original CLASSPATH? We may get ourselves into trouble if it contains some incompatible or conflict dependencies with those in IoTDB. For example, if the user has put an old Log4j into CLASSPATH, the system log may fail. Tian Jiang 发送自 Windows 10 版邮件应用 发件人: Lei Rui 发送时间: 2019年9月19日 13:42 收件人: dev@iotdb.apache.org 主题: Re: Resolve "input line is too long" in windows os Hi, Ok then it looks like `start-server.bat` needs changing because now it is ``` @REM * CLASSPATH library setting * @REM Ensure that any user defined CLASSPATH variables are not used on startup set CLASSPATH="%IOTDB_HOME%\lib" @REM For each jar in the IOTDB_HOME lib directory call append to build the CLASSPATH variable. set CLASSPATH=%CLASSPATH%;"%IOTDB_HOME%\lib\*" set CLASSPATH=%CLASSPATH%;iotdb.IoTDB goto okClasspath ``` Lei Rui On 9/19/2019 13:33,Xiangdong Huang wrote: Hi, set CLASSPATH=%CLASSPATH%;%IOTDB_CLI_HOME%\lib\* Normally, the better way is : set CLASSPATH=%IOTDB_CLI_HOME%\lib\*;%CLASSPATH% It means load %IOTDB_CLI_HOME%\lib\* first, and then %CLASSPATH% that the system original having. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Jialin Qiao 于2019年9月5日周四 上午9:21写道: Hanot iotdb-env.sh, it is in the start-client.bat, import-csv.bat and export-csv.bat... Best -- Jialin Qiao School of Software, Tsinghua University 乔嘉林 清华大学 软件学院 -原始邮件- 发件人: "Jialin Qiao" 发送时间: 2019-09-05 09:17:45 (星期四) 收件人: dev@iotdb.apache.org 抄送: 主题: Re: Resolve "input line is too long" in windows os Hi, Ok, I found this in the current iotdb-env.sh: ``` @REM Ensure that any user defined CLASSPATH variables are not used on startup set CLASSPATH="" ``` Best, -- Jialin Qiao School of Software, Tsinghua University 乔嘉林 清华大学 软件学院 -原始邮件- 发件人: "Tian Jiang" 发送时间: 2019-09-05 09:05:29 (星期四) 收件人: "dev@iotdb.apache.org" 抄送: "dev@iotdb.apache.org" 主题: Re: Resolve "input line is too long" in windows os Hi, I think it is fine, since we are not sure what %CLASSPATH% would be originally, if we depend on some unknown external dependencies, things may get uncontrollable. Tian Jiang | | Tian Jiang | | jt2594...@163.com | 签名由网易邮箱大师定制 On 9/5/2019 08:55,Jialin Qiao wrote: Hi Zesong, Thanks for sharing your experience :) I just have a question about this line: set CLASSPATH=%IOTDB_CLI_HOME%\lib\* Should it be the following command? set CLASSPATH=%CLASSPATH%;%IOTDB_CLI_HOME%\lib\* Best, -- Jialin Qiao School of Software, Tsinghua University 乔嘉林 清华大学 软件学院 -原始邮件- 发件人: "Zesong Sun" 发送时间: 2019-09-04 22:20:54 (星期三) 收件人: dev 抄送: 主题: Resolve "input line is too long" in windows os Hi, Recently I have encountered a problem when updating maven dependency in tsfile module: when starting client in windows os, I will get an error: "The input line is too long." It is resulted by limitation of the maximum length of the string that can be used at the command line: "On computers running Microsoft Windows XP or later, the maximum length of the string that you can use at the command prompt is 8191 characters. On computers running Microsoft Windows 2000 or Windows NT 4.0, the maximum length of the string that you can use at the command prompt is 2047 characters. This limitation applies to the command line, individual environment variables (such as the PATH variable) that are inherited by other processes, and all environment variable expansions. If you use Command Prompt to run batch files, this limitation also applies to batch file processing.” [1] With the suggestion of @Jialin Qiao, I reduced the dependency (from hadoop-client to hadoop-common), and local tests are all passed. However, the tests on Travis CI still failed. Thanks to @Tian Jiang, I changed the set CLASSPATH statement in .bat file of starting client: Previous: ``` for %%i in ("%IOTDB_CLI_HOME%\lib\*.jar") do call :append "%%i" goto okClasspath :append set CLASSPATH=%CLASSPATH%;%1 goto :eof ``` And now: ``` set CLASSPATH=%IOTDB_CLI_HOME%\lib\* ``` This avoid too long input line of CLASSPATH, which resolves the problem. The error confused me for really a long time... And I think maybe this solution will help you in the future. If you have any other suggestions and ideas, welcome to discuss with us : ) [1] https://support.microsoft.com/en-us/help/830473/command-prompt-cmd-exe-command-line-string-limitation BR, -- Zesong Sun School of Software, Tsinghua University 孙泽嵩 清华大学 软件学院
Re: [jira] [Created] (IOTDB-225) Support Druid connection pool
Hi, It would be nice if IoTDB supports some connection pool (e.g., Ali Druid). Does someone want to implement it? Actually, Apache Commons has also a connection pool, called Apache DBCP (though I did not never try it...) By the way, is there someone know that whether Calcite supports some connection pool? Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Jialin Qiao (Jira) 于2019年9月19日周四 上午10:32写道: > Jialin Qiao created IOTDB-225: > - > > Summary: Support Druid connection pool > Key: IOTDB-225 > URL: https://issues.apache.org/jira/browse/IOTDB-225 > Project: Apache IoTDB > Issue Type: New Feature > Reporter: Jialin Qiao > > > > > > > -- > This message was sent by Atlassian Jira > (v8.3.4#803005) >
Re: [jira] [Created] (IOTDB-223) TsFile skeleton viewer
Hi, I like the idea, and I notice the PR has been merged. A suggestion: as you consider the skeleton viewer as a tool. how about moving the java file into "tool" package, and add a script to use it? Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Lei Rui (Jira) 于2019年9月19日周四 上午12:42写道: > Lei Rui created IOTDB-223: > - > > Summary: TsFile skeleton viewer > Key: IOTDB-223 > URL: https://issues.apache.org/jira/browse/IOTDB-223 > Project: Apache IoTDB > Issue Type: Improvement > Reporter: Lei Rui > > > > > > > -- > This message was sent by Atlassian Jira > (v8.3.4#803005) >
Re: [jira] [Created] (IOTDB-222) Fix changing to read-only mode when flushing Tsfile on HDFS
Hi, I can hardly understand how to reproduce the bug (and finding why the bug appears..) I notice the issue is fixed. But my question is, is it just a bug when using HDFS? If we flush data on local disk, will the bug appear? Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Zesong Sun (Jira) 于2019年9月18日周三 下午4:45写道: > Zesong Sun created IOTDB-222: > > > Summary: Fix changing to read-only mode when flushing Tsfile > on HDFS > Key: IOTDB-222 > URL: https://issues.apache.org/jira/browse/IOTDB-222 > Project: Apache IoTDB > Issue Type: Bug > Reporter: Zesong Sun > Attachments: WechatIMG20.png > > When flushing Tsfile on HDFS, system changes to read-only mode in some > occasion. > > > > -- > This message was sent by Atlassian Jira > (v8.3.4#803005) >
Re: Resolve "input line is too long" in windows os
Hi, Ok then it looks like `start-server.bat` needs changing because now it is ``` @REM * CLASSPATH library setting * @REM Ensure that any user defined CLASSPATH variables are not used on startup set CLASSPATH="%IOTDB_HOME%\lib" @REM For each jar in the IOTDB_HOME lib directory call append to build the CLASSPATH variable. set CLASSPATH=%CLASSPATH%;"%IOTDB_HOME%\lib\*" set CLASSPATH=%CLASSPATH%;iotdb.IoTDB goto okClasspath ``` Lei Rui On 9/19/2019 13:33,Xiangdong Huang wrote: Hi, set CLASSPATH=%CLASSPATH%;%IOTDB_CLI_HOME%\lib\* Normally, the better way is : set CLASSPATH=%IOTDB_CLI_HOME%\lib\*;%CLASSPATH% It means load %IOTDB_CLI_HOME%\lib\* first, and then %CLASSPATH% that the system original having. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Jialin Qiao 于2019年9月5日周四 上午9:21写道: Hanot iotdb-env.sh, it is in the start-client.bat, import-csv.bat and export-csv.bat... Best -- Jialin Qiao School of Software, Tsinghua University 乔嘉林 清华大学 软件学院 -原始邮件- 发件人: "Jialin Qiao" 发送时间: 2019-09-05 09:17:45 (星期四) 收件人: dev@iotdb.apache.org 抄送: 主题: Re: Resolve "input line is too long" in windows os Hi, Ok, I found this in the current iotdb-env.sh: ``` @REM Ensure that any user defined CLASSPATH variables are not used on startup set CLASSPATH="" ``` Best, -- Jialin Qiao School of Software, Tsinghua University 乔嘉林 清华大学 软件学院 -原始邮件- 发件人: "Tian Jiang" 发送时间: 2019-09-05 09:05:29 (星期四) 收件人: "dev@iotdb.apache.org" 抄送: "dev@iotdb.apache.org" 主题: Re: Resolve "input line is too long" in windows os Hi, I think it is fine, since we are not sure what %CLASSPATH% would be originally, if we depend on some unknown external dependencies, things may get uncontrollable. Tian Jiang | | Tian Jiang | | jt2594...@163.com | 签名由网易邮箱大师定制 On 9/5/2019 08:55,Jialin Qiao wrote: Hi Zesong, Thanks for sharing your experience :) I just have a question about this line: set CLASSPATH=%IOTDB_CLI_HOME%\lib\* Should it be the following command? set CLASSPATH=%CLASSPATH%;%IOTDB_CLI_HOME%\lib\* Best, -- Jialin Qiao School of Software, Tsinghua University 乔嘉林 清华大学 软件学院 -原始邮件- 发件人: "Zesong Sun" 发送时间: 2019-09-04 22:20:54 (星期三) 收件人: dev 抄送: 主题: Resolve "input line is too long" in windows os Hi, Recently I have encountered a problem when updating maven dependency in tsfile module: when starting client in windows os, I will get an error: "The input line is too long." It is resulted by limitation of the maximum length of the string that can be used at the command line: "On computers running Microsoft Windows XP or later, the maximum length of the string that you can use at the command prompt is 8191 characters. On computers running Microsoft Windows 2000 or Windows NT 4.0, the maximum length of the string that you can use at the command prompt is 2047 characters. This limitation applies to the command line, individual environment variables (such as the PATH variable) that are inherited by other processes, and all environment variable expansions. If you use Command Prompt to run batch files, this limitation also applies to batch file processing.” [1] With the suggestion of @Jialin Qiao, I reduced the dependency (from hadoop-client to hadoop-common), and local tests are all passed. However, the tests on Travis CI still failed. Thanks to @Tian Jiang, I changed the set CLASSPATH statement in .bat file of starting client: Previous: ``` for %%i in ("%IOTDB_CLI_HOME%\lib\*.jar") do call :append "%%i" goto okClasspath :append set CLASSPATH=%CLASSPATH%;%1 goto :eof ``` And now: ``` set CLASSPATH=%IOTDB_CLI_HOME%\lib\* ``` This avoid too long input line of CLASSPATH, which resolves the problem. The error confused me for really a long time... And I think maybe this solution will help you in the future. If you have any other suggestions and ideas, welcome to discuss with us : ) [1] https://support.microsoft.com/en-us/help/830473/command-prompt-cmd-exe-command-line-string-limitation BR, -- Zesong Sun School of Software, Tsinghua University 孙泽嵩 清华大学 软件学院
Re: Support retention policy
Hi, I think the idea works. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Jialin Qiao 于2019年9月17日周二 上午8:58写道: > Hi, > > Each storage group has its own retention policy is fine. > In our configuration, merging can be disabled or at very low frequency. Is > it better to execute a global (bounded with the storage group) delete plan > periodically? > > Best, > -- > Jialin Qiao > School of Software, Tsinghua University > > 乔嘉林 > 清华大学 软件学院 > > > -原始邮件- > > 发件人: "Xiangdong Huang" > > 发送时间: 2019-09-17 08:32:12 (星期二) > > 收件人: dev@iotdb.apache.org > > 抄送: > > 主题: Re: [jira] [Created] (IOTDB-205) Support retention policy > > > > +1 for the feature. > > Cassandra names it as TTL. > > > > So, how about for each storage group, there is a retention policy. And, > > clean the old data when merging. > > > > --- > > Xiangdong Huang > > School of Software, Tsinghua University > > > > 黄向东 > > 清华大学 软件学院 > > > > > > Jialin Qiao (Jira) 于2019年9月9日周一 下午5:13写道: > > > > > Jialin Qiao created IOTDB-205: > > > - > > > > > > Summary: Support retention policy > > > Key: IOTDB-205 > > > URL: https://issues.apache.org/jira/browse/IOTDB-205 > > > Project: Apache IoTDB > > > Issue Type: New Feature > > > Reporter: Jialin Qiao > > > > > > > > > Many applications use TSDBs to just store recent data for a period of > > > time. For example, 10 days, and older data will be deleted > automatically. > > > Influxdb also supports this function called retention policy. > > > > > > It would be great that we also support this. > > > > > > > > > > > > -- > > > This message was sent by Atlassian Jira > > > (v8.3.2#803003) > > > >
Re: Resolve "input line is too long" in windows os
Hi, > set CLASSPATH=%CLASSPATH%;%IOTDB_CLI_HOME%\lib\* Normally, the better way is : set CLASSPATH=%IOTDB_CLI_HOME%\lib\*;%CLASSPATH% It means load %IOTDB_CLI_HOME%\lib\* first, and then %CLASSPATH% that the system original having. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Jialin Qiao 于2019年9月5日周四 上午9:21写道: > > Hanot iotdb-env.sh, it is in the start-client.bat, import-csv.bat and > export-csv.bat... > > Best > -- > Jialin Qiao > School of Software, Tsinghua University > > 乔嘉林 > 清华大学 软件学院 > > > -原始邮件- > > 发件人: "Jialin Qiao" > > 发送时间: 2019-09-05 09:17:45 (星期四) > > 收件人: dev@iotdb.apache.org > > 抄送: > > 主题: Re: Resolve "input line is too long" in windows os > > > > Hi, > > > > Ok, I found this in the current iotdb-env.sh: > > > > ``` > > @REM Ensure that any user defined CLASSPATH variables are not used on > startup > > set CLASSPATH="" > > ``` > > > > Best, > > -- > > Jialin Qiao > > School of Software, Tsinghua University > > > > 乔嘉林 > > 清华大学 软件学院 > > > > > -原始邮件- > > > 发件人: "Tian Jiang" > > > 发送时间: 2019-09-05 09:05:29 (星期四) > > > 收件人: "dev@iotdb.apache.org" > > > 抄送: "dev@iotdb.apache.org" > > > 主题: Re: Resolve "input line is too long" in windows os > > > > > > Hi, > > > > > > > > > I think it is fine, since we are not sure what %CLASSPATH% would be > originally, if we depend on some unknown external dependencies, things may > get uncontrollable. > > > > > > > > > Tian Jiang > > > > > > > > > | | > > > Tian Jiang > > > | > > > | > > > jt2594...@163.com > > > | > > > 签名由网易邮箱大师定制 > > > On 9/5/2019 08:55,Jialin Qiao wrote: > > > Hi Zesong, > > > > > > Thanks for sharing your experience :) > > > > > > I just have a question about this line: > > > > > > set CLASSPATH=%IOTDB_CLI_HOME%\lib\* > > > > > > Should it be the following command? > > > > > > set CLASSPATH=%CLASSPATH%;%IOTDB_CLI_HOME%\lib\* > > > > > > Best, > > > -- > > > Jialin Qiao > > > School of Software, Tsinghua University > > > > > > 乔嘉林 > > > 清华大学 软件学院 > > > > > > -原始邮件- > > > 发件人: "Zesong Sun" > > > 发送时间: 2019-09-04 22:20:54 (星期三) > > > 收件人: dev > > > 抄送: > > > 主题: Resolve "input line is too long" in windows os > > > > > > Hi, > > > > > > > > > > > > > > > Recently I have encountered a problem when updating maven dependency > in tsfile module: when starting client in windows os, I will get an error: > "The input line is too long." > > > > > > > > > It is resulted by limitation of the maximum length of the string that > can be used at the command line: "On computers running Microsoft Windows XP > or later, the maximum length of the string that you can use at the command > prompt is 8191 characters. On computers running Microsoft Windows 2000 or > Windows NT 4.0, the maximum length of the string that you can use at the > command prompt is 2047 characters. This limitation applies to the command > line, individual environment variables (such as the PATH variable) that are > inherited by other processes, and all environment variable expansions. If > you use Command Prompt to run batch files, this limitation also applies to > batch file processing.” [1] > > > > > > > > > With the suggestion of @Jialin Qiao, I reduced the dependency (from > hadoop-client to hadoop-common), and local tests are all passed. However, > the tests on Travis CI still failed. > > > > > > > > > Thanks to @Tian Jiang, I changed the set CLASSPATH statement in .bat > file of starting client: > > > > > > > > > Previous: > > > ``` > > > for %%i in ("%IOTDB_CLI_HOME%\lib\*.jar") do call :append "%%i" > > > goto okClasspath > > > > > > > > > :append > > > set CLASSPATH=%CLASSPATH%;%1 > > > goto :eof > > > > > > ``` > > > > > > > > > And now: > > > ``` > > > set CLASSPATH=%IOTDB_CLI_HOME%\lib\* > > > ``` > > > > > > > > > This avoid too long input line of CLASSPATH, which resolves the > problem. > > > > > > > > > The error confused me for really a long time... And I think maybe this > solution will help you in the future. If you have any other suggestions and > ideas, welcome to discuss with us : ) > > > > > > > > > > > > > > > [1] > https://support.microsoft.com/en-us/help/830473/command-prompt-cmd-exe-command-line-string-limitation > > > > > > > > > > > > > > > > > > BR, > > > -- > > > Zesong Sun > > > School of Software, Tsinghua University > > > > > > 孙泽嵩 > > > 清华大学 软件学院 >
[jira] [Created] (IOTDB-229) Inconsistent usage of Marker in TsFile
Lei Rui created IOTDB-229: - Summary: Inconsistent usage of Marker in TsFile Key: IOTDB-229 URL: https://issues.apache.org/jira/browse/IOTDB-229 Project: Apache IoTDB Issue Type: Improvement Reporter: Lei Rui -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (IOTDB-228) Catch RuntimeException in interting to memtable
Jialin Qiao created IOTDB-228: - Summary: Catch RuntimeException in interting to memtable Key: IOTDB-228 URL: https://issues.apache.org/jira/browse/IOTDB-228 Project: Apache IoTDB Issue Type: Improvement Reporter: Jialin Qiao When parsing data values into different types in the inserting process, RuntimeException often occurs and will terminate the connection. It's better to catch this and return message to users. -- This message was sent by Atlassian Jira (v8.3.4#803005)
Re: [jira] [Created] (IOTDB-181) Remove "first_value" and "last_value" in TsDigest
Hi, +1. Do not forget to log the changes into format-changelist.md Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Jialin Qiao (Jira) 于2019年8月27日周二 上午10:07写道: > Jialin Qiao created IOTDB-181: > - > > Summary: Remove "first_value" and "last_value" in TsDigest > Key: IOTDB-181 > URL: https://issues.apache.org/jira/browse/IOTDB-181 > Project: Apache IoTDB > Issue Type: Improvement > Reporter: Jialin Qiao > > > In TsDigest, we have five statistics: min_value, max_value, sum_value, > first_value, last_value. > > When filtering page or chunk, we use min_value and max_value for a filter > like "s1>10" or "s1<20". > > The sum value can be used in aggregation query "select sum(s1) from > root...". > > However, it is difficult for me to come up with a scenario that uses > "first_value" and "last_value" to filter data, and they are never used at > all... > > So why not remove them? > > > > -- > This message was sent by Atlassian Jira > (v8.3.2#803003) >
Re: [jira] [Created] (IOTDB-186) I encountered an exception when I connect to IoTDB by iotdb jdbc driver
Hi, When you saw this message, it means your server is v0.9.0-SNAPSHOT, while the JDBC is v0.8.0. Upgrading the JDBC version will solve the problem. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 邢质坦 (Jira) 于2019年8月29日周四 下午3:39写道: > 邢质坦 created IOTDB-186: > - > > Summary: I encountered an exception when I connect to IoTDB > by iotdb jdbc driver > Key: IOTDB-186 > URL: https://issues.apache.org/jira/browse/IOTDB-186 > Project: Apache IoTDB > Issue Type: Bug > Reporter: 邢质坦 > > > Exception Messge: > LF4J: Class path contains multiple SLF4J bindings. > SLF4J: Found binding in > [jar:file:/D:/Program%20Files/Apache/maven-repository/ch/qos/logback/logback-classic/1.1.11/logback-classic-1.1.11.jar!/org/slf4j/impl/StaticLoggerBinder.class] > SLF4J: Found binding in > [jar:file:/D:/Program%20Files/Apache/maven-repository/org/slf4j/slf4j-log4j12/1.7.16/slf4j-log4j12-1.7.16.jar!/org/slf4j/impl/StaticLoggerBinder.class] > SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an > explanation. > SLF4J: Actual binding is of type > [ch.qos.logback.classic.util.ContextSelectorStaticBinder] > Exception in thread "main" java.sql.SQLException: Can not establish > connection with jdbc:iotdb://10.12.20.122:6667/. > at > org.apache.iotdb.jdbc.IoTDBConnection.openSession(IoTDBConnection.java:444) > at > org.apache.iotdb.jdbc.IoTDBConnection.(IoTDBConnection.java:93) > at org.apache.iotdb.jdbc.IoTDBDriver.connect(IoTDBDriver.java:63) > at java.sql.DriverManager.getConnection(DriverManager.java:664) > at java.sql.DriverManager.getConnection(DriverManager.java:247) > at cn.edu.thu.collect.IoTDBDirectly.connect(IoTDBDirectly.java:31) > at cn.edu.thu.init.InitStorageGroup.main(InitStorageGroup.java:28) > Caused by: org.apache.thrift.protocol.TProtocolException: Required field > 'statusType' was not present! Struct: TS_Status(statusType:null) > at > org.apache.iotdb.service.rpc.thrift.TS_Status.validate(TS_Status.java:477) > at > org.apache.iotdb.service.rpc.thrift.TS_Status$TS_StatusStandardScheme.read(TS_Status.java:562) > at > org.apache.iotdb.service.rpc.thrift.TS_Status$TS_StatusStandardScheme.read(TS_Status.java:507) > at > org.apache.iotdb.service.rpc.thrift.TS_Status.read(TS_Status.java:431) > at > org.apache.iotdb.service.rpc.thrift.TSOpenSessionResp$TSOpenSessionRespStandardScheme.read(TSOpenSessionResp.java:626) > at > org.apache.iotdb.service.rpc.thrift.TSOpenSessionResp$TSOpenSessionRespStandardScheme.read(TSOpenSessionResp.java:611) > at > org.apache.iotdb.service.rpc.thrift.TSOpenSessionResp.read(TSOpenSessionResp.java:521) > at > org.apache.iotdb.service.rpc.thrift.TSIService$openSession_result$openSession_resultStandardScheme.read(TSIService.java:3151) > at > org.apache.iotdb.service.rpc.thrift.TSIService$openSession_result$openSession_resultStandardScheme.read(TSIService.java:3136) > at > org.apache.iotdb.service.rpc.thrift.TSIService$openSession_result.read(TSIService.java:3083) > at > org.apache.thrift.TServiceClient.receiveBase(TServiceClient.java:86) > at > org.apache.iotdb.service.rpc.thrift.TSIService$Client.recv_openSession(TSIService.java:156) > at > org.apache.iotdb.service.rpc.thrift.TSIService$Client.openSession(TSIService.java:143) > at > org.apache.iotdb.jdbc.IoTDBConnection.openSession(IoTDBConnection.java:419) > ... 6 more > > Process finished with exit code 1 > > > > My Java Class: > > package com.json.to.insert.web.test.demo; > > import java.sql.*; > import java.text.SimpleDateFormat; > > /** > * @author xingzhitan > * @date 2019/8/29 15:20 > */ > public class IoTDBHelloWorld { > > public static void main(String[] args) throws ClassNotFoundException, > SQLException { > Connection connection = null; > Statement statement = null; > try { > > // 1. load JDBC driver of IoTDB > Class.forName("org.apache.iotdb.jdbc.IoTDBDriver"); > // 2. DriverManager connect to IoTDB > connection = DriverManager.getConnection("jdbc:iotdb:// > 10.12.20.122:6667/", "root", "root"); > // 3. Create statement > statement = connection.createStatement(); > // 4. Set storage group > statement.execute("set storage group to root.vehicle.sensor"); > // 5. Create timeseries > statement.execute("CREATE TIMESERIES > root.vehicle.sensor.sensor0 WITH DATATYPE=DOUBLE, ENCODING=PLAIN"); > // 6. Insert data to IoTDB > statement.execute("INSERT INTO root.vehicle.sensor(timestamp, > sensor0) VALUES (2018/10/24 19:33:00, 142)"); > // 7. Query data > String sql = "select * from root.vehicl
[jira] [Created] (IOTDB-227) tsfile is stored on hdfs,but only exist storage group and time series.
邢质坦 created IOTDB-227: - Summary: tsfile is stored on hdfs,but only exist storage group and time series. Key: IOTDB-227 URL: https://issues.apache.org/jira/browse/IOTDB-227 Project: Apache IoTDB Issue Type: Bug Reporter: 邢质坦 The following exception occurred when I entered the flush command on the iotdb client. [root@bigdata-master sbin]# ./start-server.sh - Starting IoTDB - Maximum memory allocation pool = 3928MB, initial memory allocation pool = 400MB If you want to change this configuration, please check conf/iotdb-env.sh(Unix or OS X, if you use Windows, check conf/iotdb-env.bat). 2019-09-19 10:03:04,839 [main] INFO org.apache.iotdb.db.conf.IoTDBDescriptor:88 - Start to read config file ./../conf/iotdb-engine.properties 2019-09-19 10:03:04,855 [main] INFO org.apache.iotdb.db.conf.IoTDBDescriptor:108 - The stat_monitor_detect_freq_sec value is smaller than default, use default value 2019-09-19 10:03:04,856 [main] INFO org.apache.iotdb.db.conf.IoTDBDescriptor:115 - The stat_monitor_retain_interval_sec value is smaller than default, use default value 2019-09-19 10:03:04,857 [main] INFO org.apache.iotdb.db.conf.IoTDBDescriptor:222 - Time zone has been set to +08:00 2019-09-19 10:03:04,863 [main] INFO org.apache.iotdb.db.conf.IoTDBConfigCheck:54 - System configuration is ok. 2019-09-19 10:03:04,867 [main] INFO org.apache.iotdb.db.service.StartupChecks:47 - JMX is enabled to receive remote connection on port 31999 2019-09-19 10:03:04,869 [main] INFO org.apache.iotdb.db.service.StartupChecks:57 - JDK veriosn is 8. 2019-09-19 10:03:04,869 [main] INFO org.apache.iotdb.db.service.IoTDB:79 - Setting up IoTDB... 2019-09-19 10:03:04,925 [main] INFO org.apache.iotdb.db.engine.version.SimpleFileVersionController:82 - Version file updated, previous: /opt/apache-iotdb-0.9.0-SNAPSHOT-incubating/sbin/./../data/system/storage_groups/root.baic2.weldingworkshop.robotlaboratory.plc1/Version-1200, current: /opt/apache-iotdb-0.9.0-SNAPSHOT-incubating/sbin/./../data/system/storage_groups/root.baic2.weldingworkshop.robotlaboratory.plc1/Version-1300 2019-09-19 10:03:04,926 [main] INFO org.apache.iotdb.db.engine.storagegroup.StorageGroupProcessor:206 - recover Storage Group root.baic2.weldingworkshop.robotlaboratory.plc1 2019-09-19 10:03:05,045 [main] INFO org.apache.iotdb.tsfile.common.conf.TSFileDescriptor:105 - Start to read config file ./../conf/tsfile-format.properties log4j:WARN No appenders could be found for logger (org.apache.hadoop.util.Shell). log4j:WARN Please initialize the log4j system properly. log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info. 2019-09-19 10:03:06,123 [main] INFO org.apache.iotdb.db.conf.directories.DirectoryManager:80 - folder hdfs://10.12.20.131:8020/data/data/sequence doesn't exist, create it 2019-09-19 10:03:06,142 [main] INFO org.apache.iotdb.db.conf.directories.DirectoryManager:80 - folder hdfs://10.12.20.131:8020/data/data/unsequence doesn't exist, create it 2019-09-19 10:03:06,224 [main] INFO org.apache.iotdb.db.engine.storagegroup.StorageGroupProcessor:225 - root.baic2.weldingworkshop.robotlaboratory.plc1 a RecoverMergeTask root.baic2.weldingworkshop.robotlaboratory.plc1-1568858586214 starts... 2019-09-19 10:03:06,224 [main] INFO org.apache.iotdb.db.engine.merge.task.RecoverMergeTask:64 - root.baic2.weldingworkshop.robotlaboratory.plc1-1568858586214 no merge.log, merge recovery ends 2019-09-19 10:03:06,224 [main] INFO org.apache.iotdb.db.engine.StorageEngine:93 - Storage Group Processor root.baic2.weldingworkshop.robotlaboratory.plc1 is recovered successfully 2019-09-19 10:03:06,227 [main] WARN org.apache.iotdb.db.service.JMXService:109 - JMX settings in conf/iotdb-env.sh(Unix or OS X, if you use Windows, check conf/iotdb-env.bat) have been bypassed as the JMX connector server is already initialized. Please refer to iotdb-env.sh/bat for JMX configuration info 2019-09-19 10:03:06,229 [main] INFO org.apache.iotdb.db.service.JDBCService:71 - Start latch is null when getting status 2019-09-19 10:03:06,229 [main] INFO org.apache.iotdb.db.service.JDBCService:76 - Stop latch is null when getting status 2019-09-19 10:03:06,229 [main] INFO org.apache.iotdb.db.service.JDBCService:124 - IoTDB: start JDBC ServerService... 2019-09-19 10:03:06,288 [main] INFO org.apache.iotdb.db.service.JDBCService:139 - IoTDB: start JDBC ServerService successfully, listening on ip 0.0.0.0 port 6667 2019-09-19 10:03:06,292 [main] INFO org.apache.iotdb.db.cost.statistic.Measurement:142 - start measurement stats module... 2019-09-19 10:03:06,293 [main] INFO org.apache.iotdb.db.cost.statistic.Measurement:250 - start the consuming task in the measurement stats module... 2019-09-19 10:03:06,294 [main] INFO org.apache.iotdb.db.conf.adapter.ManageDy
[jira] [Created] (IOTDB-226) tsfile is stored on hdfs,we need a connector to use hive query
邢质坦 created IOTDB-226: - Summary: tsfile is stored on hdfs,we need a connector to use hive query Key: IOTDB-226 URL: https://issues.apache.org/jira/browse/IOTDB-226 Project: Apache IoTDB Issue Type: New Feature Reporter: 邢质坦 We are using Apache IoTDB to store data about industrial scenarios.tsfile is stored on HDFS.Now we need to use hive query, so I hope your development team can provide a hive connector.We want the table structure of the connector to be the same as the following. one storage group as a database; one timeseries as a table; table structure : timestamp value -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (IOTDB-225) Support Druid connection pool
Jialin Qiao created IOTDB-225: - Summary: Support Druid connection pool Key: IOTDB-225 URL: https://issues.apache.org/jira/browse/IOTDB-225 Project: Apache IoTDB Issue Type: New Feature Reporter: Jialin Qiao -- This message was sent by Atlassian Jira (v8.3.4#803005)
??????[jira] [Created] (IOTDB-224) Batch write buffer size is notcorrectly allocated
Hi~ I'm working on this issue. I will fix the size of the buffer and test it in session batch insert part. -- -- ??: "Yuan Tian (Jira)"; : 2019??9??19??(??) 9:27 ??: "dev"; : [jira] [Created] (IOTDB-224) Batch write buffer size is notcorrectly allocated Kaifeng Xue created IOTDB-224: - Summary: Batch write buffer size is not correctly allocated Key: IOTDB-224 URL: https://issues.apache.org/jira/browse/IOTDB-224 Project: Apache IoTDB Issue Type: Bug Reporter: Kaifeng Xue In RowBatch line 174, we are allocating buffer size for text datatype. But we fail to consider we need to put the length of the text firstly and then put the text itself. The bugging code as following: ``` case TEXT: for (Binary value: (Binary[]) values[i]) { valueOccupation += value.getLength(); } break; ``` -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (IOTDB-224) Batch write buffer size is not correctly allocated
Kaifeng Xue created IOTDB-224: - Summary: Batch write buffer size is not correctly allocated Key: IOTDB-224 URL: https://issues.apache.org/jira/browse/IOTDB-224 Project: Apache IoTDB Issue Type: Bug Reporter: Kaifeng Xue In RowBatch line 174, we are allocating buffer size for text datatype. But we fail to consider we need to put the length of the text firstly and then put the text itself. The bugging code as following: ``` case TEXT: for (Binary value: (Binary[]) values[i]) { valueOccupation += value.getLength(); } break; ``` -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (IOTDB-223) TsFile skeleton viewer
Lei Rui created IOTDB-223: - Summary: TsFile skeleton viewer Key: IOTDB-223 URL: https://issues.apache.org/jira/browse/IOTDB-223 Project: Apache IoTDB Issue Type: Improvement Reporter: Lei Rui -- This message was sent by Atlassian Jira (v8.3.4#803005)
Re: Release of 0.8.1
Hi, I like the idea of more releases. But I suggest to do one at a time during learning... would you do the RM this time @Xiangdong Huang? Julian Am 17.09.19, 01:02 schrieb "Xiangdong Huang" : Hi, Thanks Jialin. As there is so many bugs that be solved, how about starting to release 0.8.1? I think we can make the release more frequent to let more PPMCs know how to release. By the way, as the new API is introduced (session API), I think we can release a new version 0.9 at the same time. Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Jialin Qiao 于2019年9月17日周二 下午12:53写道: > Hi, > > I went through recent PRs and list those related to bug-fix of 0.8.0. > > * bug in updating startTime and endTime in TsFileResource > * https://github.com/apache/incubator-iotdb/pull/347 > * bug of Mix unsequence folder in sequence folder > * https://github.com/apache/incubator-iotdb/pull/393 > * bug in tsfile batch write interface > * https://github.com/apache/incubator-iotdb/pull/344 > * https://github.com/apache/incubator-iotdb/pull/385 > * bug in query with partition info of TsFile > * https://github.com/apache/incubator-iotdb/pull/382 > * bug in tsfile module in example > * https://github.com/apache/incubator-iotdb/pull/378 > * bug of the serialized size of String > * https://github.com/apache/incubator-iotdb/pull/377 > * bug of docs (maven version) > * https://github.com/apache/incubator-iotdb/pull/364 > > The 0.8.1 could contain these commits. > > Best, > -- > Jialin Qiao > School of Software, Tsinghua University > > 乔嘉林 > 清华大学 软件学院 > > > -原始邮件- > > 发件人: "Jialin Qiao (Jira)" > > 发送时间: 2019-09-17 11:53:00 (星期二) > > 收件人: dev@iotdb.apache.org > > 抄送: > > 主题: [jira] [Created] (IOTDB-216) Release of 0.8.1 > > > > Jialin Qiao created IOTDB-216: > > - > > > > Summary: Release of 0.8.1 > > Key: IOTDB-216 > > URL: https://issues.apache.org/jira/browse/IOTDB-216 > > Project: Apache IoTDB > > Issue Type: Task > > Reporter: Jialin Qiao > > > > > > This is a bug-fix version of 0.8.0 > > > > > > > > > > > > > > > > -- > > This message was sent by Atlassian Jira > > (v8.3.2#803003) >
Remind to maintain the format-changelist.md
Hi, I notice that tsfile/format-changelist.md fils is created but the file is empty.. However, there are some merged PRs changing the file format. Please help to maintain the file. It is helpful to write the data upgrade tool. I suggest to record the following info: #PR ID, changes Best, --- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院
Re: Fix changing to read-only mode when flushing Tsfile on HDFS
Hi, Changing to read-only mode may due to meeting error when flushing. It can not be solved by changing mode back, because it may happen again. Do you authorized to flush data to HDFS? -- Jialin Qiao School of Software, Tsinghua University 乔嘉林 清华大学 软件学院 > -原始邮件- > 发件人: "Zesong Sun (Jira)" > 发送时间: 2019-09-18 16:45:00 (星期三) > 收件人: dev@iotdb.apache.org > 抄送: > 主题: [jira] [Created] (IOTDB-222) Fix changing to read-only mode when flushing > Tsfile on HDFS > > Zesong Sun created IOTDB-222: > > > Summary: Fix changing to read-only mode when flushing Tsfile on > HDFS > Key: IOTDB-222 > URL: https://issues.apache.org/jira/browse/IOTDB-222 > Project: Apache IoTDB > Issue Type: Bug > Reporter: Zesong Sun > Attachments: WechatIMG20.png > > When flushing Tsfile on HDFS, system changes to read-only mode in some > occasion. > > > > -- > This message was sent by Atlassian Jira > (v8.3.4#803005)
Re: [jira] [Created] (IOTDB-221) Can I connect IoTDB with python or C?
Hi, Actually, we are working on a C client of IoTDB. Thanks to out rpc framework Thrift, this should not take too much time. We will present it soon and we would appreciate it if you can help us test it. Thanks a lot. | | Tian Jiang | | jt2594...@163.com | 签名由网易邮箱大师定制 On 9/18/2019 16:42,qinhaiqiang (Jira) wrote: qinhaiqiang created IOTDB-221: - Summary: Can I connect IoTDB with python or C? Key: IOTDB-221 URL: https://issues.apache.org/jira/browse/IOTDB-221 Project: Apache IoTDB Issue Type: Task Reporter: qinhaiqiang -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (IOTDB-222) Fix changing to read-only mode when flushing Tsfile on HDFS
Zesong Sun created IOTDB-222: Summary: Fix changing to read-only mode when flushing Tsfile on HDFS Key: IOTDB-222 URL: https://issues.apache.org/jira/browse/IOTDB-222 Project: Apache IoTDB Issue Type: Bug Reporter: Zesong Sun Attachments: WechatIMG20.png When flushing Tsfile on HDFS, system changes to read-only mode in some occasion. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Created] (IOTDB-221) Can I connect IoTDB with python or C?
qinhaiqiang created IOTDB-221: - Summary: Can I connect IoTDB with python or C? Key: IOTDB-221 URL: https://issues.apache.org/jira/browse/IOTDB-221 Project: Apache IoTDB Issue Type: Task Reporter: qinhaiqiang -- This message was sent by Atlassian Jira (v8.3.4#803005)