Prevent single failure killing playbook

Archives Ansible Project
1

Hi All,

I am currently running a playbook of several tasks, with serial:1 specified.
If any of the tasks fail on host 1, I would like to still try them on host 2 and host 3 etc. Perhaps host 1 is just too busy at the time to execute the tasks, it seems annoying to have the whole playbook fail, particularly if running it overnight.

It feels there should be a way to accomplish this, any ideas?

Thanks,
Tom

2

Hi,

I haven’t tried it my self but I think max_fail_percentage is what you are looking for.

3

Hi Omar,

Thanks for the suggestion - I tried this and was unable to get it to work. It seems correct if the value of serial is more than one, as it operates on a “per-batch” level as opposed to a “per-machine” level.

Thanks,
Tom