Open metalefty opened 1 month ago
@isudak Ansible log says, albs_test_db_1
container is started before albs_db_1
container starts. albs_db_1
container cannot be started because the same port is already used by albs_test_db_1
.
Full ansible log is here: ansible.log
time="2024-05-23T05:37:18Z" level=warning msg="/home/almalinux/albs/albs-web-server/docker-compose.yml: `version` is obsolete"
Container albs_db_1 Recreate
Container rabbitmq-ts Recreate
Container albs_sign_file_1 Recreate
Container albs_task_queue_sign_1 Recreate
Container albs_web_server_tests_1 Recreate
Container albs_immudb_1 Recreate
Container albs_redis_1 Recreate
Container albs_pulp_1 Recreate
Container albs_mosquitto_1 Recreate
Container albs_test_db_1 Recreate
Container albs_task_queue_product_modify_1 Recreate
Container albs_sign_file_1 Recreated
Container albs_db_1 Recreated
Container albs_task_queue_sign_1 Recreated
Container albs_mosquitto_1 Recreated
Container albs_gitea_listener_1 Recreate
Container rabbitmq-ts Recreated
Container albs_alts-celery_1 Recreate
Container albs_task_queue_product_modify_1 Recreated
Container albs_redis_1 Recreated
Container albs_git_cacher_1 Recreate
Container albs_web_server_tests_1 Recreated
Container albs_test_db_1 Recreated
Container albs_immudb_1 Recreated
Container albs_pulp_1 Recreated
Container albs_task_queue_errata_1 Recreate
Container albs_task_queue_1 Recreate
Container albs_task_queue_tests_1 Recreate
Container albs_web_server_1 Recreate
Container albs_task_queue_releases_1 Recreate
Container albs_task_queue_builds_1 Recreate
Container albs_gitea_listener_1 Recreated
Container albs_alts-celery_1 Recreated
Container albs_git_cacher_1 Recreated
Container albs_web_server_1 Recreated
Container albs_build_node_1 Recreate
Container albs_frontend_1 Recreate
Container albs_tests_cacher_1 Recreate
Container albs_sign_node_1 Recreate
Container albs_alts-scheduler_1 Recreate
Container albs_task_queue_1 Recreated
Container albs_task_queue_releases_1 Recreated
Container albs_task_queue_tests_1 Recreated
Container albs_task_queue_builds_1 Recreated
Container albs_task_queue_errata_1 Recreated
Container albs_sign_node_1 Recreated
Container albs_tests_cacher_1 Recreated
Container albs_alts-scheduler_1 Recreated
Container albs_frontend_1 Recreated
Container albs_nginx_1 Recreate
Container albs_build_node_1 Recreated
Container albs_nginx_1 Recreated
Container albs_mosquitto_1 Starting
Container albs_db_1 Starting
Container albs_pulp_1 Starting
Container albs_test_db_1 Starting
Container albs_sign_file_1 Starting
Container albs_web_server_tests_1 Starting
Container rabbitmq-ts Starting
Container albs_immudb_1 Starting
Container albs_redis_1 Starting
Container albs_task_queue_product_modify_1 Starting
Container albs_task_queue_sign_1 Starting
Container albs_mosquitto_1 Started
Container albs_mosquitto_1 Started
Container albs_gitea_listener_1 Starting
Container rabbitmq-ts Started
Container albs_alts-celery_1 Starting
Container albs_task_queue_sign_1 Started
Container albs_test_db_1 Started
Container albs_task_queue_product_modify_1 Started
Container albs_web_server_tests_1 Started
Container albs_immudb_1 Started
Container albs_redis_1 Started
Container albs_sign_file_1 Started
Container albs_pulp_1 Started
Container albs_gitea_listener_1 Started
Container albs_alts-celery_1 Started
Error response from daemon: driver failed programming external connectivity on endpoint albs_db_1 (62970255b998ea4128b98da92d188af912817f7f3b0f072c1c518ca14e392016): Bind for 0.0.0.0:5432 failed: port is already allocated
Starting every service found in docker-compose.yml
doesn't look right. Maybe all images should be created, but only albs_web_server_1
has to be started?
Starting every service found in
docker-compose.yml
doesn't look right. Maybe all images should be created, but onlyalbs_web_server_1
has to be started?
docker compose up
will bring everything on docker-compose.yml
up.
In my case, I rarely do that when dealing with my dev env, and I don't like the idea of albs-deploy bringing up every service, but never wrote an issue or talked to anybody about this (shame on me). I'd say that we should only bring up the minimal viable system, as an example:
Then, ensure that we have a good documentation about it in albs-deploy or in our wiki. As for db, redis, etc, they will be started as they are dependencies of some of the services above. I believe that the machinery for that is already in albs-deploy, but is not used by default. What do you think?
Thanks, then I think excluded_containers
should be filled by default to start all necessary containers out-of-box.
It might be better making a list of necessary containers rather than a list of containers to exclude.
It might be better making a list of necessary containers rather than a list of containers to exclude.
The logic used around excluded_containers
can be found here. It looks solid to me, but I don't have a strong opinion on whether we should change it or not.
After https://github.com/AlmaLinux/albs-web-server/pull/792, there is a problem.
albs_test_db_1 container fails to start due to already used port 5432/tcp.
test_db
service added by https://github.com/AlmaLinux/albs-web-server/pull/792 listens 5432/tcp despite the port is already used bydb
service.