Hello friends, I’ve noticed that the cluster I recently created is very slow when inserting data using the “insert” command. Upon checking the logs, it seems that the ETL phase of the job is taking a considerable amount of time. I would like to know how to fix this issue. I’m planning to use the “Insert” command for data insertion as a test. My table structure is simple, and there is currently no data in the table.
What is concerned slow from your point of view and why are you only inserting 1 record? StarRocks is an OLAP engine NOT an OLTP engine.
I think the second OLAP was supposed to be OLTP?
No… sub-second and second OLAP IS NOT an OLTP; it’s just a faster OLAP. Is StarRocks an OLAP or OLTP database?. Sub-second and second OLAP are not optimized for single record inserts however they can deal with streaming inserts and upserts through micro batching (eg. commit n around of records every 5 seconds). If done correctly, from the initial insertation of the kafka (can use others) record and making it queryable in a sql query to StarRocks is in mid to high single digit seconds.
Could you post a profile about the insert in this place?
I tested the writing using the ‘Insert’ method. If it’s just this mechanism, then I’ll test it using another method. Thank you.
So the insert speed is very odd so we really need more info. Even on my laptop, I get insert of 1 record in 400ms. It seems like you’re using a load method but I cannot tell what it is. Without all the info we ask, we cannot give you good advice.
I was reading the statement another way. You’re correct. The second OLAP was supposed to be OLTP. I fixed the original post.
