I’ve many windows packer jobs that have been running for many months and even years in some cases.
We’ve always had the odd winrm issue where it never connects but normally rerunning the packer job works without any adjustment. However, over the last week. our success rates have dropped to around 1 in 8 runs being successful. I’ve historically used the winrm enablement example code from https://www.packer.io/docs/communicators/winrm
but noticed this week that there is a new recommendation to not use quick config and use a different approach
Using the exact copy of the latter script hasn’t yet shown any improvement in success rates. I’m using the windows 2016 base image from MS and the standard AWS ebs builder configuration. I’ve also tried the latest packer version. Before I start a deep dive on one of the failing image I’d like to ask if anybody has experienced similar issues or maybe converted to using ssh and found it more reliable? I’d really welcome your thoughts.