Containers created with AddContainer(..).WithLifetime(..)

[raffaeler]

I have a solution where the container is added with something like this:

var vectordb = builder.AddContainer("vectorDB", "qdrant/qdrant")
    .WithEndpoint(6333, 6333, "http", "dashboard")
    .WithEndpoint(6334, 6334, null, "grpc")
    .WithBindMount(
        @"H:\VMs\ContainerData\qdrant\storage",
        "/qdrant/storage")
    .WithBindMount(
        @"H:\VMs\ContainerData\qdrant\config\production.yaml",
        "/qdrant/config/production.yaml")
    .WithBindMount(
        @"H:\VMs\ContainerData\qdrant\snapshots",
        "/qdrant/snapshots")
    .WithLifetime(ContainerLifetime.Persistent);

I love the ContainerLifetime.Persistent as it saves me a lot of time between runs.
Anyway, the container is only usable when the solution is running: as soon as you shut down the application the container is still running but the ports are "removed" and the qdrant dashboard cannot be reached anymore.

From Docker Destkop, when the Aspire solution is stopped I see the following:

• The vibrant_joloit container was created by myself, it's obviously stopped but you can clearly see the port mappings being persisted
• The vectorDB-db4a9ca4 was created by Aspire, it is still running because of the ContainerLifetime.Persistent but has no port mappings and cannot be reached

If I run the Aspire solution again, the qdrant dashboard can be reached again.

Is there a way for Aspire to persist the port and volume settings persisted so that it can be always reached?

[davidfowl]

Are you trying to use persistent containers as an easy way to setup a container that is always running on your machine? There might be a mismatch on the expected use case.

[raffaeler]

Nope. You still need to verify the entries in the database or doing some maintenance, backup, etc. when the Aspire solution is not running.
I do not expect needing to run the entire Aspire solution (8+ processes) to just open the db dashboard and delete a collection. Some maintenance operations may fail when the solution is running.

Also, there are times where you may need to run a project alone, without all the other projects. But if you don't run Aspire, you can't access the database.

The only possible alternative is temporary comment out AddContainer and remembering to manually run the container first, but this is not very practical because you also have to keep in sync the endponts and volumes so that they are available in the Aspire manifest.