Description of problem: NVMe-cli 2.x breaks NVMeOF connection at least with TCP protocol, when issuing connect-all command. Manual connection (discover target NQN and target IP addresses, use single controller connect commands) works. Version-Release number of selected component (if applicable): libnvme-1.4 Steps to Reproduce: 1. example: nvme connect-all -t tcp -a 172.168.25.64 -s 4420 Actual results: Failed to write to /dev/nvme-fabrics: Invalid argument failed to connect controller, error 1006 Failed to write to /dev/nvme-fabrics: Invalid argument failed to connect controller, error 1006 Failed to write to /dev/nvme-fabrics: Invalid argument failed to connect controller, error 1006 Failed to write to /dev/nvme-fabrics: Invalid argument failed to connect controller, error 1006 dmesg: [ 8045.246675] nvme_fabrics: unknown parameter or missing value 'tls' in ctrl creation request [ 8045.246707] nvme_fabrics: unknown parameter or missing value 'tls' in ctrl creation request [ 8045.246719] nvme_fabrics: unknown parameter or missing value 'tls' in ctrl creation request [ 8045.246730] nvme_fabrics: unknown parameter or missing value 'tls' in ctrl creation request Expected results: Additional info:
Hello If you provide the package you want me to test I can test it. I already validated the fix from Chris Leech so I am actually happy for you to just say tested by CEE ~~~ From the other bug This is a known issue The TLS issue, I already reported in the other case is well known now and we have a fix. We have it working now with no issues with a test package I got from Chris Leech. https://github.com/linux-nvme/libnvme/issues/612 Run rpm -Uvh libnvme-1.2-2.1.test.el9_2.x86_64.rpm ~~~