-
Notifications
You must be signed in to change notification settings - Fork 543
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
It seems to me that a connection pool is being created and shut down for every transaction in NiFi. #1817
Comments
Can you update to the latest version of the JDBC driver |
@com-exem-route7 thank you for reporting the issue and providing so many details! |
I tried applying |
After appling Could there be an issue with the ClickHouse connection settings in NiFi?
|
I've also looked at the heap dump, but since NiFi was recently restarted, there isn't much significant information yet. I expect that after about a week, with Full Garbage Collection (FGC), we should be able to see updates. I’d like to reduce the heap size and increase throughput to get faster results, but the data generated by the pipeline is being used by other teams, so I can't make those changes. I will provide updates as soon as there's new information. |
@com-exem-route7 thank you for so detailed information! |
|
OK. I will use shaded jar until the issue you shared is resolved.
Thank you for sharing the issue. I think I need to test the parameters of the jdbc url I set.
I set up the connection by referring to this document. I will also test it without the load balancing parameter(s). Until next Wednesday, Korea will be observing a holiday period, so it might be difficult to conduct tests or updates. Thank you for reviewing this issue. |
@chernser
I didn't quite understand this statement. For reference, I am using Thank you again. 😊 |
Good day, @com-exem-route7 ! When the client is reused - the internal connection pool would work. But it is true only for Apache Http Client as the provider. |
@chernser
|
Good day, @com-exem-route7 ! Thanks! |
Hello, @chernser I now understand what you were talking about above. I will test with Thank you for kindly explaining it. |
Hello, @chernser . I applied the configuration to the development NiFi and monitored it alongside the production NiFi for comparison. Configuration (Dev)
Observations:
Thank you. |
Description
Hello Team.
I am using NiFi and clickhouse-jdbc to collect data and load it into ClickHouse.
However, over time, NiFi's JVM Old Generation memory usage keeps increasing.
After generating and analyzing a heap dump, I suspect that the issue might be related to clickhouse-jdbc.
Based on the logs, it seems to me that a connection pool is being created and shut down for every transaction.
Could you help me determine if this is a problem with NiFi or with clickhouse-jdbc?
JConsole G1 Old Generation Heap
Spec
Heap Dump (Eclipse MAT)
Leak Suspects
Dominator Tree
Leak Hunter
+)
I enabled debug logs for the
com.clickhouse.client
andcom.clickhouse.jdbc
packages in NiFi. I have a question: theConnection pool shut down
log is being printed for every transaction. Is this normal behavior? It doesn't seem like the connections in the connection pool are being reused.PoolingHttpClientConnectionManager Connection pool shut down
was logged 6125 times over a period of 20 seconds.The text was updated successfully, but these errors were encountered: