mattsre opened a new issue, #269:
URL: https://github.com/apache/pulsar-client-python/issues/269

   While running Pulsar cluster proxies behind an AWS Load Balancer I ran into 
issues with TLS connections to the proxies using example code from this 
repository. Running the `examples/consumer.py` code with the client url set to 
an LB hostname and `use_tls=True` results in connection failures, whereas 
simply specifying `pulsar+ssl://` the consumer works as expected. After 
checking docs it seems the `use_tls` argument is deprecated, but I would still 
expect the behavior between `use_tls` and `pulsar+ssl://` to be the same. I can 
provide details on the load balancer configuration if needed.
   
   Examples with some logs (fake hostname):
   ```python
   client = pulsar.Client("pulsar://pulsar-primary.acme.org:6651", use_tls=True)
   ```
   This client produces connection errors like below:
   ```
   2025-10-10 12:56:28.815 INFO  [0x1f7cd60c0] Client:86 | Subscribing on Topic 
< redacted_topic >
   2025-10-10 12:56:28.815 INFO  [0x1f7cd60c0] ClientConnection:193 | [<none> 
-> pulsar://pulsar-primary.acme.org:6651] Create ClientConnection, timeout=10000
   2025-10-10 12:56:28.815 INFO  [0x1f7cd60c0] ConnectionPool:124 | Created 
connection for 
pulsar://pulsar-primary.acme.org:6651-pulsar://pulsar-primary.acme.org:6651-0
   2025-10-10 12:56:28.875 INFO  [0x16cf07000] ClientConnection:410 | 
[192.168.100.214:65139 -> < redacted_ip >] Connected to broker
   2025-10-10 12:56:38.836 ERROR [0x16cf07000] ClientConnection:634 | 
[192.168.100.214:65139 -> < redacted_ip >] Connection was not established in 
10000 ms, close the socket
   2025-10-10 12:56:38.837 INFO  [0x16cf07000] ClientConnection:1336 | 
[192.168.100.214:65139 -> < redacted_ip >] Connection disconnected (refCnt: 2)
   2025-10-10 12:56:38.837 INFO  [0x16cf07000] ConnectionPool:141 | Remove 
connection for 
pulsar://pulsar-primary.acme.org:6651-pulsar://pulsar-primary.acme.org:6651-0
   2025-10-10 12:56:38.837 INFO  [0x16cf07000] RetryableOperation:114 | 
Reschedule get-partition-metadata-persistent:/< redacted_topic > for 100 ms, 
remaining time: 29900 ms
   2025-10-10 12:56:38.837 INFO  [0x16cf07000] ClientConnection:282 | 
[192.168.100.214:65139 -> < redacted_ip >] Destroyed connection to 
pulsar://pulsar-primary.acme.org:6651-0
   ```
   
   After switching the client to use `pulsar+ssl://`, the example consumer code 
works with no other changes.
   ```python
   client = pulsar.Client("pulsar+ssl://pulsar-primary.acme.org:6651")
   ```
   
   Is this difference in behavior known and/or expected?
   


-- 
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]

Reply via email to