This is the error i get:

"options": {
      "headers": {
        "user-agent": "Renovate Bot (GitHub App 2740)",
        "accept-encoding": "gzip, deflate",
        "authorization": "***********",
        "accept": "application/json"
      },
      "url": "https://proget.inedo.com/v2/productimages/inedo/proget/tags/list?n=10000",
      "username": "",
      "password": "",
      "method": "GET",
      "http2": false
    },
    "response": {
      "statusCode": 500,
      "statusMessage": "Internal Server Error",
      "body": null,
      "headers": {
        "cache-control": "public, max-age=300",
        "content-type": "application/json; charset=utf-8",
        "content-encoding": "gzip",
        "vary": "Authorization",
        "server": "Microsoft-IIS/10.0",
        "x-proget-version": "5.3.27.18",
        "docker-distribution-api-version": "registry/2.0",
        "x-aspnet-version": "4.0.30319",
        "x-powered-by": "ASP.NET",
        "date": "Wed, 21 Apr 2021 05:57:33 GMT",
        "connection": "close",
        "content-length": "84"
      },
      "httpVersion": "1.1"
    }
  }

Seeing 500 on https://proget.inedo.com/containers/registry?feedId=22.

Maybe that is related?

I know, i saw that difference. Just wanted to inform you

It's now always dooing upgrade

ProGet version is 5.3.24.7.
Current DB schema version is 5.3.24.
Executing untracked script DDL-DML/0000. Untracked/10. Block 5.3 Upgrades.sql...
Executing untracked script DDL-DML/0000. Untracked/10. Block SQL Server 2005.sql...
Tracked script DDL-DML/v1-3/0002. 0.1.0/10 SET DATABASE PROPERTIES.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/20 CREATE TYPES.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/21 CREATE ROLE ProGetUser_Role.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30 CREATE TABLE Configuration.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30 CREATE TABLE Environments.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30CTC1 CREATE TABLE Connectors.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30CTC2 CREATE TABLE ConnectorFilters.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30EVT1 CREATE TABLE EventTypes.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30EVT2 CREATE TABLE EventTypeDetails.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30EVT3 CREATE TABLE EventOccurences.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30EVT4 CREATE TABLE EventOccurenceDetails.sql has already been executed; skipping...
Tracked script DDL-DML/v1-3/0002. 0.1.0/30USR1 CREATE TABLE Groups.sql has already been executed; skipping...

It should skip updates update at all if already uptodate, otherwise startup will be very slow

@viceice said in [ProGet] Manual database upgrade (docker, kubernetes):

upgradedb

needs to be ProGet.Service upgradedb. fixed it in my pr

getting this when trying to run upgrade command

For help on a specific command, type: ProGet.Service help [command]
To run interactively, type: ProGet.Service interactive
Invalid command: upgrade

ProGet.Service 5.3.24.7
Usage: ProGet.Service [command]

Commands:
run Runs the ProGet service and/or the ProGet web server interactively.
install Installs the ProGet service as a Windows service.
uninstall Uninstalls the ProGet Windows service.
resetadminpassword Switches to the built-in user directory and changes the Admin account password to "Admin".
upgradedb Upgrades the database schema version to this version of ProGet.

For help on a specific command, type: ProGet.Service help [command]
To run interactively, type: ProGet.Service interactive

https://github.com/Inedo/inedo-docs/pull/223

The issue is closed, so can i implement it to my kubernetes deployment?

ok, database upgrade command can be run in init_container (wicch can be without any probes)

here a new sample:

kind: StatefulSet
apiVersion: apps/v1
metadata:
  name: proget
  namespace: services
  labels:
    app: proget
spec:
  replicas: 1
  selector:
    matchLabels:
      app: proget
  template:
    metadata:
      labels:
        app: proget
    spec:
      volumes:
        - name: packages
          emptyDir: {} # should be pvc
        - name: extensions
          emptyDir: {} # should be pvc
      init_container:
        - name: upgrade-db
          image: 'ghcr.io/visualon/proget:5.3.21'
          command: ["/usr/local/proget/service/ProGet.Service.exe", "upgrade"] # Updates the database if required
          env:
            - name: SQL_CONNECTION_STRING
              value: Data Source=proget-sql; Initial Catalog=ProGet; User ID=sa; Password=<YourStrong!Passw0rd> # should be come from secret
          volumeMounts:
            - name: packages
              mountPath: /var/proget/packages
            - name: extensions
              mountPath: /var/proget/extensions
      containers:
        - name: proget
          image: 'ghcr.io/visualon/proget:5.3.21'
          ports:
            - containerPort: 80
              protocol: TCP
          env:
            - name: SQL_CONNECTION_STRING
              value: Data Source=proget-sql; Initial Catalog=ProGet; User ID=sa; Password=<YourStrong!Passw0rd> # should be come from secret
          volumeMounts:
            - name: packages
              mountPath: /var/proget/packages
            - name: extensions
              mountPath: /var/proget/extensions
          livenessProbe: # additional startupProbe and readyness probes are available
            httpGet:
              path: /log-in
              port: 80
            initialDelaySeconds: 15
            periodSeconds: 5
  serviceName: proget

This will call the database update command with an init container which has no timeouts. After that is done (exit code zero), kubernetes will start the normal container(s) which can skip upgrade, as it's aready done.

So an proget upgrade would simply be an image update

probes: https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-startup-probes

So the init container runs before the normal proget container.

Normally the proget container gets some startup and liveness probes, if they fail proget will be restartet.

So if i upgrade the image and a database update is needed, this would currently trap theed by the probes and i get an endless restart loop.

Of cause i can extend the startup probes timeout, but that would cause a lot of failues in kubernetes logs.