I’m trying out accelerated mode on a bunch of Centos6.4 nodes and getting random “decryption failed” errors on a few nodes, while many others work fine.
Here’s /var/log/messages on an afflicted node:
https://gist.github.com/davidbirdsong/7366779
I’ve seen that a couple of times. Killing the python process on the affected nodes and re-running will fix it. I haven’t been able to reliably reproduce this so far. Generally it happens to me when an ansible-playbook run fails in the middle.
This might be a timing issue with the rekey.
Please open a github ticket with some steps to reproduce this (as best you can).