uninstall.sh --remove-data today only runs docker compose down -v (containers + volumes) -- it explicitly leaves everything else behind and tells the customer to clean up manually:
"Configuration files remain at ${INSTALL_DIR} -- remove that directory yourself if you want a completely clean slate."
It also never removes the pulled argus-{api,ui,proxy} (and third-party timescaledb/valkey/docker-socket-proxy) images, the argus-network Docker network install.sh creates, or runs docker system prune. A customer who genuinely wants ARGUS fully gone from a box has to know all of this and do it by hand.
Expected behavior
A new option, e.g. ./uninstall.sh --purge (on top of the existing --remove-data, kept as-is for the "remove data but I might reinstall" case), that additionally:
docker compose down -v --remove-orphans
Removes any leftover argus-* named containers
Removes the argus-{api,ui,proxy} images (and the third-party base images, clearly warning first since those may be shared with other stacks on the same host)
Removes argus-* named volumes
Removes the argus-network network
Removes the install directory itself ($INSTALL_DIR)
docker system prune -f -- with a clear warning that this affects any unused Docker resource on the host, not just ARGUS's own, since a shared host might have other stacks
Step 7 in particular needs a loud, explicit confirmation separate from the existing --remove-data confirmation, since docker system prune is host-wide, not ARGUS-scoped -- this should probably require its own explicit --yes acknowledgment or a second typed confirmation, not be silently bundled into --purge.
Priority
Low -- convenience/completeness for uninstall, not a bug or security issue. Genuinely useful for support/testing scenarios (a clean-slate reinstall) though.
Affected repository/components
argus-installer -- uninstall.sh.
Related: this is installer lifecycle tooling, part of the broader roadmap tracked centrally in weboria/argus-appliance#22.
## Problem
`uninstall.sh --remove-data` today only runs `docker compose down -v` (containers + volumes) -- it explicitly leaves everything else behind and tells the customer to clean up manually:
> "Configuration files remain at ${INSTALL_DIR} -- remove that directory yourself if you want a completely clean slate."
It also never removes the pulled `argus-{api,ui,proxy}` (and third-party `timescaledb`/`valkey`/`docker-socket-proxy`) images, the `argus-network` Docker network `install.sh` creates, or runs `docker system prune`. A customer who genuinely wants ARGUS fully gone from a box has to know all of this and do it by hand.
## Expected behavior
A new option, e.g. `./uninstall.sh --purge` (on top of the existing `--remove-data`, kept as-is for the "remove data but I might reinstall" case), that additionally:
1. `docker compose down -v --remove-orphans`
2. Removes any leftover `argus-*` named containers
3. Removes the `argus-{api,ui,proxy}` images (and the third-party base images, clearly warning first since those may be shared with other stacks on the same host)
4. Removes `argus-*` named volumes
5. Removes the `argus-network` network
6. Removes the install directory itself (`$INSTALL_DIR`)
7. `docker system prune -f` -- with a clear warning that this affects *any* unused Docker resource on the host, not just ARGUS's own, since a shared host might have other stacks
Step 7 in particular needs a loud, explicit confirmation separate from the existing `--remove-data` confirmation, since `docker system prune` is host-wide, not ARGUS-scoped -- this should probably require its own explicit `--yes` acknowledgment or a second typed confirmation, not be silently bundled into `--purge`.
## Priority
Low -- convenience/completeness for uninstall, not a bug or security issue. Genuinely useful for support/testing scenarios (a clean-slate reinstall) though.
## Affected repository/components
argus-installer -- `uninstall.sh`.
Related: this is installer lifecycle tooling, part of the broader roadmap tracked centrally in weboria/argus-appliance#22.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Problem
uninstall.sh --remove-datatoday only runsdocker compose down -v(containers + volumes) -- it explicitly leaves everything else behind and tells the customer to clean up manually:It also never removes the pulled
argus-{api,ui,proxy}(and third-partytimescaledb/valkey/docker-socket-proxy) images, theargus-networkDocker networkinstall.shcreates, or runsdocker system prune. A customer who genuinely wants ARGUS fully gone from a box has to know all of this and do it by hand.Expected behavior
A new option, e.g.
./uninstall.sh --purge(on top of the existing--remove-data, kept as-is for the "remove data but I might reinstall" case), that additionally:docker compose down -v --remove-orphansargus-*named containersargus-{api,ui,proxy}images (and the third-party base images, clearly warning first since those may be shared with other stacks on the same host)argus-*named volumesargus-networknetwork$INSTALL_DIR)docker system prune -f-- with a clear warning that this affects any unused Docker resource on the host, not just ARGUS's own, since a shared host might have other stacksStep 7 in particular needs a loud, explicit confirmation separate from the existing
--remove-dataconfirmation, sincedocker system pruneis host-wide, not ARGUS-scoped -- this should probably require its own explicit--yesacknowledgment or a second typed confirmation, not be silently bundled into--purge.Priority
Low -- convenience/completeness for uninstall, not a bug or security issue. Genuinely useful for support/testing scenarios (a clean-slate reinstall) though.
Affected repository/components
argus-installer --
uninstall.sh.Related: this is installer lifecycle tooling, part of the broader roadmap tracked centrally in weboria/argus-appliance#22.