最近测试 titan 性能，部署了 3 个节点的 tikv 集群，单独部署 titan，然后压测时发现，相同的数据，每个节点平均 30+ region 时，GET qps 大概 7w。调整 region 数后，每个 节点的 region 为1300+个，再次用同样的命令压测，GET qps 能达到 13w，请问 titan 是否有根据 region 数做并发控制呢？

Question

Currently, there are too mach structures contain various informer/control/lister, I think we can define a dedicated structure to store all shared informer/control/lister, like Dependences in Kubernetes to avoid pass the parameters everywhere.

Currently, the JDBC string containing the host name and port is hardcoded. For example, consider the connection string for MySQL:

https://github.com/sqlancer/sqlancer/blob/e62224edcb43fe421dc87601f75090739e34e595/src/sqlancer/mysql/MySQLProvider.java#L248

It would be desirable to make this configurable. We could provide an option to set the JDBC string for each DBMS. For example, for MySQL,

Please answer these questions before submitting your issue. Thanks!

1. What did you do?
   If possible, provide a recipe for reproducing the error.
   #317 said that we support restore empty database and table, but for now, empty databases won't be backup at all!

Let we do it:

• Make an empty database;
• Backup it by BR;
• Drop it, and restore.

2. What did you expect to see?
   The emp