You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@iotdb.apache.org by GitBox <gi...@apache.org> on 2020/06/28 04:47:31 UTC
[GitHub] [incubator-iotdb] kqkdChen opened a new issue #1425: SQL执行时间怎样优化到毫秒级别？还是说根本就不可能？
kqkdChen opened a new issue #1425:
URL: https://github.com/apache/incubator-iotdb/issues/1425
data count: **18999994** total
SQL: **select * from root.position where time > 1559179819000 limit 10**
执行上述sql：
shell client显示花费时间： **6.171S**
IOTDB Metrics Server：**0-7ms??????????**
为什么两者差别如此之大？
我还尝试使用jdbc api，时间花费在5s，我以为查询会很强，没想到这么弱，感觉比不上HBase...
这是你们的产品介绍，我不知道是否是我的用法有问题，查询根本做不到毫秒级别，超过300W数据就已经是秒级别了，官网的文档也不详细，基本的性能调优我都没有找到。
求解惑，还有尽快完善文档吧，连集群模式都没有，生产谁敢用...
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-iotdb] kqkdChen commented on issue #1425: SQL执行时间怎样优化到毫秒级别？还是说根本就不可能？
Posted by GitBox <gi...@apache.org>.
kqkdChen commented on issue #1425:
URL: https://github.com/apache/incubator-iotdb/issues/1425#issuecomment-650695760
CPU Cores: 8 Total, 29% CPU Ratio
JVM Memory: 1969 1969 1295 (Max/Total/Free)MB
Host Memory: 16 GB Total, 5.0 GB Used
18999994条数据 count 下竟然将近9秒。
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-iotdb] qiaojialin commented on issue #1425: SQL执行时间怎样优化到毫秒级别？还是说根本就不可能？
Posted by GitBox <gi...@apache.org>.
qiaojialin commented on issue #1425:
URL: https://github.com/apache/incubator-iotdb/issues/1425#issuecomment-653715782
了解了一下，有两个原因。
第一个是数据写入是并行写一个时间段的数据，导致磁盘上相互重叠的乱序数据块非常多，查一个点也需要读所有重叠的数据块做 merge，乱序也会影响聚合查询，本来可以读预计算的值，有乱序就需要读原始数据。
第二个是我们一个默认参数设置的不太好，avg_series_point_number_threshold 设置的比较，之前默认是10000，序列少、序列比较长的时候，数据块就比较小，对查询不友好。现在改成了 30万。
客户端注意一下写入模式，只要顺序写就没问题了。乱序可以有，但是太多会影响性能。
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-iotdb] jixuan1989 commented on issue #1425: SQL执行时间怎样优化到毫秒级别？还是说根本就不可能？
Posted by GitBox <gi...@apache.org>.
jixuan1989 commented on issue #1425:
URL: https://github.com/apache/incubator-iotdb/issues/1425#issuecomment-653711224
> select * from root.position where time > 1559179819000 limit 10
如果limit10的话，不应该6s吧？ @qiaojialin
> 18999994条数据 count 下竟然将近9秒。
请问这个有乱序数据吗？ （data/data/unsequence 目录是否未空？）
另外，目前主分支还没合并PISA索引，#1439
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-iotdb] kqkdChen closed issue #1425: SQL执行时间怎样优化到毫秒级别？还是说根本就不可能？
Posted by GitBox <gi...@apache.org>.
kqkdChen closed issue #1425:
URL: https://github.com/apache/incubator-iotdb/issues/1425
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org