ferenc-csaky commented on PR #23183:
URL: https://github.com/apache/flink/pull/23183#issuecomment-1694598625
> I run this locally, it makes sense to add this improvement. It would be
great if we could make it explicit in the logs (maybe adding an extra log line)
that the following is for the data bind port:
>
> ```
> 2023-08-26 21:30:45,211 INFO
org.apache.flink.runtime.io.network.netty.NettyConfig [] - NettyConfig
[server address: localhost/127.0.0.1, server port range: 55000-55100, ssl
enabled: false, memory segment size (bytes): 32768, transport type: AUTO,
number of server threads: 1 (manual), number of client threads: 1 (manual),
server connect backlog: 0 (use Netty's default), client connect timeout (sec):
120, send/receive buffer size (bytes): 0 (use Netty's default)]
> 2023-08-26 21:30:45,246 INFO
org.apache.flink.runtime.io.network.NettyShuffleServiceFactory [] - Created a
new FileChannelManager for storing result partitions of BLOCKING shuffles. Used
directories:
>
/var/folders/yh/t9bt8gwj4zsd949jnr55vx980000gn/T/flink-netty-shuffle-9c342614-dc5f-467f-9a9d-1736246bdc62
> 2023-08-26 21:30:45,280 INFO
org.apache.flink.runtime.io.network.buffer.NetworkBufferPool [] - Allocated 128
MB for network buffer pool (number of memory segments: 4096, bytes per segment:
32768).
> 2023-08-26 21:30:45,289 INFO
org.apache.flink.runtime.io.network.NettyShuffleEnvironment [] - Starting the
network environment and its components.
> 2023-08-26 21:30:45,309 INFO
org.apache.flink.runtime.io.network.netty.NettyClient [] - Transport
type 'auto': using NIO.
> 2023-08-26 21:30:45,310 INFO
org.apache.flink.runtime.io.network.netty.NettyClient [] - Successful
initialization (took 20 ms).
> 2023-08-26 21:30:45,312 INFO
org.apache.flink.runtime.io.network.netty.NettyServer [] - Transport
type 'auto': using NIO.
> 2023-08-26 21:30:45,341 INFO
org.apache.flink.runtime.io.network.netty.NettyServer [] - Successful
initialization (took 30 ms). Listening on SocketAddress /127.0.0.1:55000.
> ```
I'm not sure where you think it would be appropriate. I think Netty specific
classes are too general to add that info explicitly, what can be done easily is
to highlight the listening port explicitly when the connection is successfully
established (last line is the newly added log):
```
2023-08-27 09:42:16,255 INFO
org.apache.flink.runtime.taskexecutor.TaskManagerRunner [] - Starting
TaskManager with ResourceID: localhost:59695-945f3f
2023-08-27 09:42:16,286 INFO
org.apache.flink.runtime.taskexecutor.TaskManagerServices [] - Temporary
file directory '/var/folders/wp/ccy48gw1255bswh9bx9svxjc0000gn/T': total 460
GB, usable 125 GB (27.17% usable)
2023-08-27 09:42:16,287 INFO
org.apache.flink.runtime.io.disk.iomanager.IOManager [] - Created a new
FileChannelManager for spilling of task related data to disk (joins, sorting,
...). Used directories:
/var/folders/wp/ccy48gw1255bswh9bx9svxjc0000gn/T/flink-io-6c07c533-5a9f-42e2-abeb-a3b35042bd57
2023-08-27 09:42:16,291 INFO
org.apache.flink.runtime.io.network.netty.NettyConfig [] - NettyConfig
[server address: localhost/127.0.0.1, server port range: 55000-55100, ssl
enabled: false, memory segment size (bytes): 32768, transport type: AUTO,
number of server threads: 1 (manual), number of client threads: 1 (manual),
server connect backlog: 0 (use Netty's default), client connect timeout (sec):
120, send/receive buffer size (bytes): 0 (use Netty's default)]
2023-08-27 09:42:16,324 INFO
org.apache.flink.runtime.io.network.NettyShuffleServiceFactory [] - Created a
new FileChannelManager for storing result partitions of BLOCKING shuffles. Used
directories:
/var/folders/wp/ccy48gw1255bswh9bx9svxjc0000gn/T/flink-netty-shuffle-2e98bd72-3306-4ddc-b551-3d3c77d95cfa
2023-08-27 09:42:16,345 INFO
org.apache.flink.runtime.io.network.buffer.NetworkBufferPool [] - Allocated 128
MB for network buffer pool (number of memory segments: 4096, bytes per segment:
32768).
2023-08-27 09:42:16,352 INFO
org.apache.flink.runtime.io.network.NettyShuffleEnvironment [] - Starting the
network environment and its components.
2023-08-27 09:42:16,374 INFO
org.apache.flink.runtime.io.network.netty.NettyClient [] - Transport
type 'auto': using NIO.
2023-08-27 09:42:16,374 INFO
org.apache.flink.runtime.io.network.netty.NettyClient [] - Successful
initialization (took 21 ms).
2023-08-27 09:42:16,376 INFO
org.apache.flink.runtime.io.network.netty.NettyServer [] - Transport
type 'auto': using NIO.
2023-08-27 09:42:16,395 INFO
org.apache.flink.runtime.io.network.netty.NettyServer [] - Successful
initialization (took 19 ms). Listening on SocketAddress /127.0.0.1:55000.
2023-08-27 09:42:16,395 INFO
org.apache.flink.runtime.taskexecutor.TaskManagerServices [] - TaskManager
connection initialized successully; listening internally on port: 55000
```
The other option would be to log it before the process IMO, but since the
whole thing is started by the `TaskManagerRunner`, it would be more like
redundant information beforehand. Tying the actual internal listening port to
the TM explicitly can be useful and make things moe obvious.
WDYT?
--
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]
