Kyperr opened a new issue, #4508:
URL: https://github.com/apache/arrow-adbc/issues/4508
### What happened?
The Flight SQL driver's FlightSqlConnection seeds its primary client into an
internal
Caffeine cache (clientCache) that is configured with expireAfterAccess(5,
MINUTES) and a
removal listener that closes evicted clients. The primary client is the same
object the
connection uses directly for createStatement(), getInfo(), getObjects(), etc.
Any workload that only talks to the coordinator never touches the cache, so
after 5 idle
minutes the primary entry expires and the removal listener closes the shared
primary
client. From then on every operation on the connection fails, even though
the caller never
closed it.
Additionally, this functionality prevents any use-cases that include pooling
these objects.
### Stack Trace
_No response_
### How can we reproduce the bug?
I've attached a file that reproduces the issue.
[CacheReproducer.java](https://github.com/user-attachments/files/29987334/CacheReproducer.java)
### Environment/Setup
_No response_
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
