• Bug
  • Status: Closed
  • 2 Major
  • Resolution: Won't Fix
  • drb
  • Reporter: hhuynh
  • November 10, 2007
  • 0
  • Watchers: 0
  • November 12, 2007
  • November 12, 2007

Description

There’s no guaranteed server process will start before client processes when we run TC Maven plugin. When clients end up starting first, user has to wait for 2 minutes before clients retry. This is mostly a usability issue.

[INFO] [consumer] 2007-11-10 17:26:58,467 WARN - Connection refused from server: 192.168.1.100:9510 [INFO] [node3] 2007-11-10 17:26:58,567 WARN - Connection refused from server: 192.168.1.100:9510 [INFO] [node2] 2007-11-10 17:26:58,788 WARN - Connection refused from server: 192.168.1.100:9510 [INFO] [producer] 2007-11-10 17:26:59,088 WARN - Connection refused from server: 192.168.1.100:9510 [INFO] [consumer] 2007-11-10 17:27:04,468 WARN - Connection refused from server: 192.168.1.100:9510 [INFO] [node3] 2007-11-10 17:27:04,568 WARN - Connection refused from server: 192.168.1.100:9510 [INFO] [node2] 2007-11-10 17:27:04,788 WARN - Connection refused from server: 192.168.1.100:9510 [INFO] [producer] 2007-11-10 17:27:05,088 WARN - Connection refused from server: 192.168.1.100:9510 [INFO] [dso start] 2007-11-10 17:27:05,990 INFO - Starting reconnect window: 120000 ms. [INFO] [dso start] 2007-11-10 17:27:06,017 INFO - Terracotta Server has started up as ACTIVE node on port 9510 successfully, and is now ready for work.

Comments

Steve Harris 2007-11-10

I don’t understand why that would be? Is this a case where we are running in persistent mode and all the clients and all the servers are stopped and started again leading to the new server waiting the full reconnect time to be able to reconnect?

BTW, please follow fiona’s template on bug reports. This is kind of vague.

Hung Huynh 2007-11-10

No this is nothing to do with persistence or restart. As a client start up, it tries to connect to server so many times (2 times as the above log suggested) If server isn’t available around that time, it waits for 2 minutes before retry.

It doesn’t happen all the time and there’s no consistent way to reproduce. Just try tc-maven-plugin-sample under “examples” of the maven plugin repo.

Hung Huynh 2007-11-12

With Tim and Orion help, I have figured this out why this is the case. I have persistence mode on and upon rerun my project with “mvn tc:run”, if in previous run, the server exited before the clients, it’s gonna remember this in the next run and enforce a 2 minutes reconnect window for all clients.

I didn’t see this all the time because sometimes I ran it with “mvn clean tc:run” and the persisted data was wiped out.

I’m not sure if this is an issue after understanding this is how things suppose to work.