okay i have only played with harvestor a long time ago, but i was really interested in this particul issue (because i was board) and the MAC part. there were some previous instances of that error in a previous version of harvester, where it looked like it was an issue with the actual harvester-rancher pods themselves. similar issue in the past https://github.com/harvester/harvester/issues/8001 for some ideas https://support.switcherstudio.com/article/322-security-error-bad-record-message-authentication-code-mac is it possible that some sort of time drift is happening and pushing things out of the server time/tls comfort range? I recently had an issue (or noticed) that one of my certificate servers (non rancher) had a crazy date. Because my main services (step-ca) were not waiting for the NTP to start/sync before starting.