[BUG] up: an optional (required: false) service_healthy dependent is not started when a required dependent shares the failing wait
Los mantenedores suelen responder en 1 día
Evaluación
- Dificultad
- 4/5
- Tiempo estimado
- 3-5 días
- Aptitud para principiantes
- 58/100
Línea de trabajo
Start by tracing how the start phase handles the shared service_healthy wait and how abortReason decides whether each dependent is blocked; the issue points to the plan engine change in #14081 and a possible connection to #14311. Reproduce with the supplied compose file and up -d db needs_db optional_db, then compare with a run that leaves out needs_db. Done when the required dependent fails while optional_db starts with the optional dependency warning.
Escrito por el modelo de indexación a partir del texto del issue.
Descripción
Description
With v5.7.1, up does not start a service whose service_healthy dependency is optional (required: false) when another service in the same run depends on that dependency as required and the dependency turns unhealthy. The optional dependent is given up together with the required one, with the event:
not started: condition "service_healthy" of service "db" not met
When the required dependent is not part of the run, the same optional dependent is started as expected (with the optional dependency "db" failed to start warning). So whether an optional dependent starts depends on whether some other service needs the same dependency as required.
v5.6.0 starts the optional dependent in both cases.
Steps to reproduce
A compose file along these lines (a dependency whose healthcheck always fails, one required and one optional service_healthy dependent):
services:
db:
image: busybox
command: sleep infinity
healthcheck:
test: ["CMD-SHELL", "exit 1"]
interval: 2s
timeout: 3s
retries: 1
needs_db:
image: busybox
command: sleep infinity
depends_on:
db:
condition: service_healthy
optional_db:
image: busybox
command: sleep infinity
depends_on:
db:
condition: service_healthy
required: false
Then up -d db needs_db optional_db.
Expected
needs_db fails (its required condition is not met); optional_db is started anyway, with the optional-dependency warning, as v5.6.0 does:
Container db Error dependency db failed to start
Container db Skipped: optional dependency "db" failed to start
level=warning msg="optional dependency \"db\" failed to start: container db is unhealthy"
Container optional_db Started
dependency failed to start: container db is unhealthy
Actual (v5.7.1)
optional_db is not started; its row gets not started: condition "service_healthy" of service "db" not met. With needs_db left out of the run, optional_db starts.
Notes
From reading the code, the service_healthy wait on db looks like a single plan node shared by both dependents. When it fails (with required semantics, for needs_db), it cancels the phase, and abortReason reports optional_db as held by "the very wait that failed" (unmetCondition(...) + " not met"), although for optional_db that wait should not block its start. This seems to have come with the start phase moving into the plan engine (#14081). Possibly related to #14311 (start -p treating an optional dependency as required).
Versions
- Seen with the Go SDK
github.com/docker/compose/v5v5.7.1, callingUp(the reproduction above was run through an application embedding it, with real images rather than busybox; the file above is the minimal equivalent). - The
docker composeCLI v5.6.0 on the same host and project starts the optional dependent. - Not tried with the v5.7.1 CLI binary.
- Docker Engine on Linux.
- Lenguaje dominante
- Go
- Estrellas
- 38.9k
- Forks
- 6k
- Merge medio
- 2 d 22 h
- PR fusionados (30 d)
- 81
Preparar el entorno
- Incluye un Dockerfile o un archivo de Docker Compose
- Tiene una plantilla de pull request
- Leer la guía de contribución
Primeros pasos
- Lee el issue completo y luego la guía de contribución del proyecto.
- Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
- Haz un fork del repositorio y trabaja en una rama.
- Abre un pull request que haga referencia al número del issue.
Más de docker/compose
-
[BUG] `docker compose port` prints `:0` (exit 0) for a port the image exposes but the service does not publishPosiblemente ocupada @ndeloof la tomó hace 1 día. Abierto
Dificultad 2/5 1-3 horas Aptitud para principiantes 78/100
Los mantenedores suelen responder en 1 día
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 62/100
docker/compose#14435 · 2 comentarios ·
Los mantenedores suelen responder en 1 día
-
[BUG] deploy.restart_policy without condition drops the restart policyQuizá libre de nuevo @akshat-lakhera la tomó hace 25 días y no hay ningún pull request abierto. Abierto
Dificultad 2/5 1-3 horas Aptitud para principiantes 82/100
docker/compose#14383 · 3 comentarios ·
Los mantenedores suelen responder en 1 día
-
[BUG] Negative replica count in scale flags (docker compose scale / --scale) is accepted without validationPosiblemente ocupada @arsalanyavari la tomó hace 8 días. Abiertokind/bug status/approved
Dificultad 2/5 1-3 horas Aptitud para principiantes 30/100
Los mantenedores suelen responder en 1 día
-
[BUG] pre_stop hooks are skipped when up or scale stops a containerPosiblemente ocupada @htoyoda18 la tomó hace 11 días. Abiertokind/bug status/approved
Dificultad 4/5 3-5 días Aptitud para principiantes 53/100
Los mantenedores suelen responder en 1 día
Todos los issues de docker/compose
Issues similares
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 78/100
IBM-Cloud/terraform-provider-ibm#7115 ·
Los mantenedores suelen responder en 2 días
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 86/100
fluxcd/notification-controller#1441 ·
Los mantenedores suelen responder en 1 día
-
bug
Dificultad 2/5 1-3 horas Aptitud para principiantes 66/100
volcengine/volcengine-go-sdk#163 · 1 comentario ·
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 76/100
OneBusAway/maglev#2145 ·
Los mantenedores suelen responder en 2 días
-
priority/medium status/ready type/chore
Dificultad 2/5 1-3 horas Aptitud para principiantes 84/100
discobox-ai/discobox#1357 ·
Los mantenedores suelen responder en 1 día