Somewhat recently we’ve upgraded our Octopus instance and after, we started seeing sporadic errors with connectivity during health checks and/or deployments (Linux targets):
The screenshot doesn’t reflect it but the errors seemed to be occurring only with servers where Calamari was being upgraded dynamically before the deployment. In order to restore the client to a healthy status, we had to disable the original tentacle instance an re-register the server.
Even after the upgrade we’re getting reports that occasionally there are failed deployments to healthy servers and the only “fix” is to re-register the target. Any ideas on what the issue could be?
Since the issue is intermittent, is it possible you have multiple SSH targets with the same Fingerprint? You can confirm this with a simple SQL select query:
SELECT COUNT(DISTINCT Fingerprint) as Unique_Fingerprints from dbo.Machine
SELECT COUNT(DISTINCT Id) as Unique_Machine_Ids from dbo.Machine WHERE Fingerprint is not NULL
Let me know your thoughts at your earliest convenience.
If I’m interpreting the output correctly we’ve got 341 duplicate fingerprints out there. Thanks for pointing me in the right direction. I’ll reach out again if I run into any more road blocks.