Overseerr → Seerr¶

Upstream Overseerr has merged with Jellyseerr into a new project: Seerr. The overseerr-k8s charm is in maintenance mode and will be removed in a future release. This guide walks you through moving your existing Overseerr deployment to Seerr without losing requests, users, or settings.

What you get:

Plex, Jellyfin, and Emby support (Overseerr is Plex-only)
Active upstream - bug fixes, security patches, new features
Identical Radarr/Sonarr integration via Juju relations
All existing requests, users, and settings preserved

The migration is explicit and user-driven. Nothing changes until you run the steps below. Seerr can run alongside Overseerr during cutover - disable Overseerr only when you're satisfied Seerr is healthy.

Before you start¶

Back up your Overseerr config PVC. This is the migration's only destructive step; a snapshot or manual kubectl cp of /config to your laptop gives you a rollback path.
Confirm your overseerr-k8s revision includes the export-config action. Run juju run overseerr/0 export-config --help. The action is only available on latest/edge — track 1 stable doesn't have it. If your overseerr is on 1/stable (or any other channel), refresh first:
```
juju refresh overseerr --channel=latest/edge
```
juju refresh is non-destructive for channel switches: the app stays alive and keeps its data, you just get a newer charm revision.

Quick Deploy users following this flow will hit this automatically: bumping your bundle ?ref=track/1 → ?ref=main (and running terraform apply) refreshes overseerr to the main bundle's default latest/edge channel.

Identify pod and model names. This guide uses <model> and <overseerr-pod> / <seerr-pod> placeholders. Substitute your actual values:

juju models                    # find <model>
juju status overseerr          # pod is <app>-<unit>, e.g. "overseerr-0"

Warning

juju scp is broken for K8s sidecar containers on Juju 3.6.x - the tar stream silently drops mid-transfer and the destination file ends up empty even though the command exits 0. This guide uses kubectl cp for all file copies.

1. Export the Overseerr config¶

juju run overseerr/0 export-config

The action tars /config (excluding logs/, cache/, *.tmp) to /config/overseerr-export.tgz inside the workload container. Output:

copy-command: |
  kubectl -n <model> cp -c overseerr <overseerr-pod>:/config/overseerr-export.tgz
    ./overseerr-export.tgz
path: /config/overseerr-export.tgz
sha256: 5d1b46df5cc7a7ecd8b8922d35c0891ddc8646edf3f9852f1ecb5c77657d57f2
size: "8333"

Save the sha256 - you'll use it to verify the transfer.

2. Pull the tarball¶

kubectl -n <model> cp -c overseerr <overseerr-pod>:/config/overseerr-export.tgz ./export.tgz
sha256sum ./export.tgz
# should match the sha256 from step 1

3. Enable Seerr alongside Overseerr¶

Quick DeployManual Deploy

Add enable_seerr = true to your main.tf:

module "charmarr" {
  # ... your existing config ...

  enable_overseerr = true   # keep running until cutover
  enable_seerr     = true   # new
}

Apply:

terraform apply

The bundle deploys seerr-k8s and creates all relations (Radarr/Sonarr, Plex, ingress, service-mesh). Wait for the Seerr unit to reach waiting: Complete setup in web UI - that's the post-first-start state.

Warning

Do not complete the web UI setup yet - the next step will overwrite the config.

juju deploy seerr-k8s --trust --channel=latest/edge seerr

juju integrate seerr:media-manager radarr:media-manager
juju integrate seerr:media-manager sonarr:media-manager
juju integrate seerr:media-server  plex:media-server

# optional: dedicated ingress (mirrors your existing overseerr-ingress)
juju deploy istio-ingress-k8s --trust --channel=dev/edge seerr-ingress
juju integrate seerr:istio-ingress-route seerr-ingress:istio-ingress-route

# optional: service mesh
juju integrate seerr:service-mesh beacon:service-mesh

Wait for juju status seerr to show waiting: Complete setup in web UI. Don't complete the wizard.

4. Push the tarball to Seerr¶

kubectl -n <model> cp -c seerr ./export.tgz <seerr-pod>:/app/config/import.tgz

5. Run `import-config`¶

juju run seerr/0 import-config \
  path=/app/config/import.tgz \
  sha256=<sha256-from-step-1>

The action stops the Seerr workload, wipes /app/config, extracts the tarball, fixes ownership, and replans. On next start, Seerr's upstream auto-migration rewrites the Overseerr schema into Seerr's format. A backup of the original settings.json is preserved inside the workload container as /app/config/settings.old.json.

6. Verify¶

juju status seerr

Once Seerr finishes its first start with the imported config, the unit will sit at waiting: Complete setup in web UI again. This is normal - Seerr regenerates session secrets on the schema upgrade, so you need to re-validate the Plex OAuth in the web UI.

Sanity-check the data:

Open the Seerr UI via your ingress (or port-forward).
Sign in with your Plex account.
Settings → Services - Radarr and Sonarr should be pre-configured.
Users - existing accounts should be listed.
Requests - history should be intact.

7. Decommission Overseerr¶

Only do this once you've verified Seerr is healthy and serving requests.

Quick DeployManual Deploy

Set enable_overseerr = false in your main.tf:

module "charmarr" {
  # ... your existing config ...

  enable_overseerr = false  # decommission
  enable_seerr     = true
}

terraform apply

The bundle removes the overseerr app, the overseerr-ingress helper, and all their relations.

juju remove-application overseerr
juju remove-application overseerr-ingress   # if you ran a dedicated ingress

The Overseerr PVC will be released. Hang onto your backup tarball for a few days as insurance.

Troubleshooting¶

import-config fails with Tarball not found at /app/config/import.tgz

The kubectl cp in step 4 didn't land the file. Verify:

kubectl -n <model> exec <seerr-pod> -c seerr -- ls -la /app/config/import.tgz

If missing, re-run the kubectl cp. Make sure you pass -c seerr - without it, kubectl cp copies into the charm container, not the workload.

import-config fails with sha256 mismatch

The tarball was corrupted in transit, or you passed the wrong checksum. Re-run export-config, re-copy, and pass the new checksum. The sha256 parameter is optional - you can omit it to skip verification, but it's strongly recommended.

Seerr won't start after import

Check the workload logs:

kubectl -n <model> logs <seerr-pod> -c seerr

Common causes: a corrupt database file in the tarball, or filesystem ownership wrong. The import-config action runs chown -R 1000:1000 /app/config, but if your storage backend has restrictive permissions you may need to fix them manually.

I lost data - how do I roll back?

Stop Seerr (enable_seerr = false + terraform apply, or juju remove-application seerr), restore the Overseerr PVC from your backup if it was destroyed, and you're back where you started. Because the migration is action-driven and Overseerr is only read from (never modified), rolling back is just "use the backup."

Can I migrate without downtime?

Not cleanly. Both apps would need to share Plex/Radarr/Sonarr state, and Seerr's first-start auto-migration is destructive to its own /app/config. Plan a brief maintenance window. The actual data transfer takes seconds; most of the time is Seerr restarting.

What changes after migration¶

	Overseerr	Seerr
OCI image	`lscr.io/linuxserver/overseerr`	`ghcr.io/seerr-team/seerr`
Config path	`/config`	`/app/config`
Port	5055	5055
Media servers	Plex only	Plex, Jellyfin, Emby
API endpoints	`/api/v1/...`	`/api/v1/...` (compatible)
Settings file	`settings.json`	`settings.json` (+ `settings.old.json` post-migration)
Relations on the charm	`media-manager`, `media-server`, `istio-ingress-route`, …	same
Juju app name	`overseerr` (convention)	`seerr` (convention)

The /api/v1/ API is backwards-compatible - any external integrations (mobile apps, browser extensions, third-party tools) keep working with the new endpoint URL.