Upgrading via abra app upgrade doesn't undeploy services/apps that were removed from the recipe #931

New Issue

2026-09-15T13:44:58Z

simon commented

2026-09-15 13:44:58 +00:00

When upgrading our nodes from the old to the new monitoring stack I noticed that the old containers remained running on the node. Think this is an abra issue?

After the upgrade docker service ls listed these

sl7zdspv6603   monitoring_XYZ_kolli_cloud_app         replicated   1/1        grafana/alloy:v1.18.1
tw1ai0wledr5   monitoring_XYZ_kolli_cloud_cadvisor    replicated   1/1        gcr.io/cadvisor/cadvisor:v0.49.1
nvtflg8eo12f   monitoring_XYZ_kolli_cloud_promtail    replicated   1/1        grafana/promtail:2.9.9

When upgrading our nodes from the old to the new [monitoring stack](https://git.coopcloud.tech/coop-cloud/monitoring-ng) I noticed that the old containers remained running on the node. Think this is an abra issue? After the upgrade `docker service ls` listed these ``` sl7zdspv6603 monitoring_XYZ_kolli_cloud_app replicated 1/1 grafana/alloy:v1.18.1 tw1ai0wledr5 monitoring_XYZ_kolli_cloud_cadvisor replicated 1/1 gcr.io/cadvisor/cadvisor:v0.49.1 nvtflg8eo12f monitoring_XYZ_kolli_cloud_promtail replicated 1/1 grafana/promtail:2.9.9 ```

decentral1se added the bug label 2026-10-19 08:58:18 +00:00

decentral1se commented

2026-10-19 08:58:24 +00:00

That would have to be the swarm runtime not figuring things out. Abra describes what should happen and it's swarms job to make that happen. I've seen this happen elsewhere, especially in the case of many containers. I believe you could reproduce this with the Docker CLI also because swarm is just not handling itself well in some cases. If you can provide a "smoking gun" report with logs or more info on abra definitely causing this, then please re-open the issue!

That would have to be the swarm runtime not figuring things out. Abra describes what should happen and it's swarms job to make that happen. I've seen this happen elsewhere, especially in the case of many containers. I believe you could reproduce this with the Docker CLI also because swarm is just not handling itself well in some cases. If you can provide a "smoking gun" report with logs or more info on `abra` definitely causing this, then please re-open the issue!

decentral1se closed this issue

2026-10-19 08:58:25 +00:00

Sign in to join this conversation.

2 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: toolshed/abra#931

Upgrading via abra app upgrade doesn't undeploy services/apps that were removed from the recipe #931

Upgrading via `abra app upgrade` doesn't undeploy services/apps that were removed from the recipe #931