tests: internal: timeout: added coroutine hang test case

[leonardo-albertovich]

This unit test shoult be used to verify the validity of the proposed fixes for the issue that arises when a connection attempt times out, the fixes are :

.1 - PR #4140 - Resumes the coroutines in the timeout handler
.2 - https://github.com/fluent/fluent-bit/tree/leonardo-event-injection-poc - Injects a synthesized event to the event loop to naturally cause the process to be awoken

Signed-off-by: Leonardo Alminana [email protected]

[leonardo-albertovich]

There's a strange error in some Ubuntu 18.04 virtual machines which is odd but unrelated to the patch, this is a dump of the output :

[2022/01/12 16:49:18] [ info] [storage] version=1.1.5, initializing...
[2022/01/12 16:49:18] [ info] [storage] in-memory
[2022/01/12 16:49:18] [ info] [storage] normal synchronization mode, checksum disabled, max_chunks_up=128
[2022/01/12 16:49:18] [ info] [cmetrics] version=0.2.2
[2022/01/12 16:49:18] [ info] [sp] stream processor started
Error: 1/12 16:49:19] [error] [net] TCP connection failed: 128.1.1.1:65534 (Connection refused)
Error: 1/12 16:49:19] [error] [output:tcp:tcp.0] no upstream connections available to 128.1.1.1:65534
[2022/01/12 16:49:19] [ warn] [engine] failed to flush chunk '20196-1642006159.5863193.flb', retry in 9 seconds: task_id=0, input=dummy.0 > output=tcp.0 (out_id=0)
[2022/01/12 16:49:28] [ warn] [engine] service will shutdown in max 30 seconds
Error: 1/12 16:49:28] [error] [net] TCP connection failed: 128.1.1.1:65534 (Connection refused)
Error: 1/12 16:49:28] [error] [output:tcp:tcp.0] no upstream connections available to 128.1.1.1:65534
[2022/01/12 16:49:29] [ info] [task] dummy/dummy.0 has 1 pending task(s):
[2022/01/12 16:49:29] [ info] [task]   task_id=0 still running on route(s): tcp/tcp.0 
[2022/01/12 16:49:58] [ info] [task] dummy/dummy.0 has 1 pending task(s):
[2022/01/12 16:49:58] [ info] [task]   task_id=0 still running on route(s): tcp/tcp.0 
[2022/01/12 16:49:58] [ info] [engine] service has stopped (1 pending tasks)
[ FAILED ]
  timeout.c:92: Check for hung coroutines... failed
FAILED: 1 of 1 unit tests has failed.

What I see there is that :

1. The connection attempt to 128.1.1.1 is refused which is odd considering this worked on other machines and suggests it might not be related to the endpoint but the virtual machine
2. A second connection seems to be attempted and that's probably because Retry_limit is set to 1 instead of no_retries

I will change the retry_limit setting to no_retries but we will probably have to set up our own blackholed endpoint for this test to reliably work.

[edsiper]

thanks.

Please fix the commits subjects typo so we can merge it:

tests: interal:...

[leonardo-albertovich]

Fixed the typo but please don't merge it yet, we need to change the endpoint information, I already sent a request to have one hosted by us as this public one is not reliable.

[leonardo-albertovich]

This is ready to be merged @edsiper

[edsiper]

thanks