Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.
Original topic: 开启tls的tidb集群如何动态添加tispark组件
[Test Environment for TiDB] Testing Environment
The version is 5.4. How can I dynamically expand TiSpark in an existing cluster (with TLS authentication enabled)?
I used the scale-out.yaml format and executed tiup cluster scale-out tidb-test scale-out.yaml, but I got the following message:
How can I add the TiSpark machine to the existing cluster?
Additional information:
There is an existing independent Spark cluster (version 3.3)
The TiSpark package is: tispark-assembly-3.3_2.12-3.2.0.jar
Do not use tiup to install Spark in production; the Spark installed by tiup does not have high availability.
Uh, it’s just a test environment to verify something… How do you expand it, boss? The old cluster already has data in it.
The error indicates that TiSpark does not support TLS.
Is there an official explanation? Is it not supported to add dynamically? Shutting down the cluster, configuring it, and then restarting it would also work.
It looks like we need to upgrade to 3.0, and the Spark configuration files need to be adjusted accordingly. I haven’t done it before, so I’m not sure. Please research it and share your findings.
TiSpark is using the latest version 3.2.0. I’ve searched for a long time but couldn’t find an answer. Let’s see if any other experts know…
I tried it, but it still doesn’t work…
I see it on the official website.
But I haven’t tried it…
This configuration is for Spark, and adding TiSpark to the existing TiDB cluster is not the same thing, right?
I recommend manually deploying Spark and TiSpark.
TiSpark is actually just a jar package, simply a “bridge” for Spark to access TiKV/TiFlash.
Yes, the current issue is that after configuring Spark, I encountered an access error. I believe it’s because the topology of my TiDB cluster hasn’t been configured with TiSpark. If that’s not the case, could you please advise? (So, can I use Spark directly without configuring TiSpark on the TiDB side?)
The Spark error message is as follows:
No, it needs to be configured.
My test environment # did not enable TLS
When using it, use
tidb_catalog first:
TiSpark 用户指南 | PingCAP 文档中心
Otherwise, it defaults to querying Hive, if I remember correctly.
There is no problem with the query.
However, I tried it without using cat, and it seems to work as well. 
The question then goes back to how to expand TiSpark nodes on a cluster with TLS authentication enabled. 
Tomorrow I will study it in my environment
|
username: WRAllen
|
username: 
