Unstable 1.34.0 on ECS Fargate

[Abhinab-AY]

We accidentally upgraded to 1.34.0 on staging ECS cluster and that gave us headache(s). When there's any kind of response or error on the underlying PHP application, it kills the unitd service and in return stops exposing port 80.

• This wasn't the case on 1.33.0.
• There were no config changes.
• Simple container re-build introduced this issue

Observed that when one of the thread dies , it doesn't spawn a new one. Resulting in port 80 going down. There is not indication on what is the source of this error. Example of the error right before the container crash.

""(process 1512) exited with code 0""

Workaround - Installed 1.33.0 to hard code the version.

Let me know if you need more details.

[ac000]

Observed that when one of the thread dies , it doesn't spawn a new one. Resulting in port 80 going down. There is not

Do you mean one of unit's router threads?

""(process 1512) exited with code 0""

That looks like a normal exit condition.

Workaround - Installed 1.33.0 to hard code the version.

Let me know if you need more details.

If you have a coredump, even a minimal backtrace from that would be useful.

It's hard to tell, but I wonder if this and this are related to this and this which is hopefully addressed by this.

[ac000]

1.34.1 probably addresses this. I'll go ahead and close this, feel free to re-open if 1.34.1 still has issues...