We used to call docker ... log from GitHub Action workflows as a follow-up step when the main step in the workflow fails. However, recently we switched to using docker compose to bring up storage backends, and our integration scripts do docker compose down as a teardown step in the script itself. This means that calling docker compose .. logs from the workflow afterwards does nothing, since the containers are gone. The "right" way to do this is like this: https://github.com/jaegertracing/jaeger/blob/9a30dfc15df4e6712cf02344acae18256781153a/scripts/cassandra-integration-test.sh#L30-L43
Task: apply the above approach to other scripts as needed and clean up their corresponding workflows.
We used to call
docker ... log
from GitHub Action workflows as a follow-up step when the main step in the workflow fails. However, recently we switched to usingdocker compose
to bring up storage backends, and our integration scripts dodocker compose down
as a teardown step in the script itself. This means that callingdocker compose .. logs
from the workflow afterwards does nothing, since the containers are gone. The "right" way to do this is like this: https://github.com/jaegertracing/jaeger/blob/9a30dfc15df4e6712cf02344acae18256781153a/scripts/cassandra-integration-test.sh#L30-L43Task: apply the above approach to other scripts as needed and clean up their corresponding workflows.