Debugging Running SQL Queries with Control Center
One of the issues you can encounter when running GridGain is a long query execution. You can see this from logs, but Control Center provides extra tools to help you localize the issue.
When debugging the issue, navigate to the SQL tab and its Running queries screen. Here you can stop the long running queries to restore cluster functionality. This will not last, as the same issue is likely to happen again, but will allow your service to continue working while you look for a solution.
After that you can switch to the Query Statistics screen. Here you can see which queries execute consistently slow. Click Explain to dive deeper into query execution process.
In distributed systems, queries are executed in 2 phases:
Map - invoked on each node to process data locally;
Reduce - invoked on the query coordinator node and aggregate data from other nodes.
For example, one of possible reasons is that SQL uses
scan instead of
index. In this case, you can create a secondary index and speed up query execution dramatically.