Private Packagist Self-Hosted update requires KOTS upgrade

If the dashboard shows:

'This version of Private Packagist Self-Hosted requires a version of KOTS that is different than what you currently have installed.'

Please refer to the "Updates" section in our maintenance guide.

Generating a support bundle

You can generate a support bundle from the Replicated Management Console on port 8800 by navigating to the Troubleshoot tab, clicking "Generate a support bundle", and selecting "Analyze". Once the analysis is done, either download the bundle and manually send it to us or click on the send icon which will send us the bundle. Please always notify us if you send us a support bundle!

Issues with Multi-factor Authentication

If you are having problems setting up MFA, or are unable to log in via MFA, with your generated codes, there may be a time-drift issue with either the Self-hosted Private Packagist server or the device you are using to generate the codes.

To make sure that the Self-hosted Private Packagist server is synchronized to the correct time, you should check that both the current server time and timezone are set to the correct values. If you can enable Network Time Protocol (NTP) for the server, we also recommend doing that.

The methods for doing so will vary depending on the underlying server Operating System.

Please be aware that offline-based TOTP hardware can drift up to a few minutes a year. As Private Packagist only allows a time-drift of up to one (1) minute, we recommend using TOTP devices that have the ability to stay synchronized with the correct time (such as a phone, or re-programmable TOTP hardware devices).

Issues with Reverse-Proxy running in front of the Kubernetes Cluster

Please follow the instructions below, if you are experiencing problems with the reverse-proxy not being able to connect to the cluster and encountering errors like this:

Peer closed connection in SSL handshake (104: Connection reset by peer) while SSL handshaking to upstream

The following examples assume you are using nginx as a reverse-proxy. Please consult the documentation of other reverse-proxy servers to achieve the same result.

Ensure that the SNI (Server Name Indication) TLS Extension is properly set for requests to the ingress controller of the Kubernetes Cluster. This is not the case when using IPs in proxy_pass and will result in an SSL handshake error.

To pass the SNI hostname from the incoming request to the upstream server, add the following directives to nginx:

proxy_ssl_name $host;
proxy_ssl_server_name on;

If you are using different hostnames on the upstream and on the reverse-proxy, set the value in the proxy_ssl_name directive to the corresponding hostname of the upstream server.

Issues after changing the Private Packagist Self-Hosted domain name

If you've changed the domain name used to access your Private Packagist Self-Hosted Kubernetes installation, you'll need to clear the Composer endpoint Redis cache.

Important: Please note that there are multiple databases created in your Redis instance, and that Redis is not solely used for caching. Be careful not to flush any of the other databases.

If you are not using the built-in Redis database, check your internal documentation how to connect to the Redis instance.

kubectl exec -it redis-0 -- redis-cli -p 9869
select 5
flushdb

Afterwards, run composer update mirrors to make sure all repository references are up-to-date.