When Ansible fails with -vvv, it gives me a big dump of debug info that basically says “I copied an ansible script to the system and ran it”, with all the SSH details and whatnot I might need if I was trying to debug the process of copying and running the script.
My problem is that most of the time the problem wasn’t with getting the script over there or executing it, but an unexpected failure during execution, and I get precious little information about that. What I want is for Ansible to be able to also tell me “here are the commands this script had run when it failed”, kind of like running a bash script with bash -x.
Without that, the most helpful thing I get when a failure occurs is (to use my most recent issue as an example) tried to run usr/bin/git ls-remote origin -h refs/heads/master but then failed with $ERROR. But to really troubleshoot this I also need to know things like the command that was used to clone the repository (or however origin was defined), etc. I could SSH into the machine and poke around, but unless the failure was catastrophic it seems that the tmp environment in which the script was run is deleted, so I can’t even do that.
Any suggestions for digging deeper into problems like this? (not the specific git issue, but Ansible failures in general)
Thanks!
–Brad