Fail fast if ansible-init failed #608
Conversation
| - name: Check ansible-init hasn't failed (yet) | ||
| # NB: only allows early exit if it has, does not catch future failures! | ||
| assert: | ||
| that: "'failed' not in _ansible_init_failed.stdout" |
There was a problem hiding this comment.
[rocky@cclr-dev-svn3-dr08-u14 ~]$ sudo systemctl status ansible-init
● ansible-init.service
Loaded: loaded (/etc/systemd/system/ansible-init.service; enabled; vendor preset: disabled)
Active: activating (auto-restart) (Result: exit-code) since Fri 2025-03-07 17:37:16 UTC; 1s ago
Process: 24805 ExecStart=/usr/bin/ansible-init (code=exited, status=1/FAILURE)
Main PID: 24805 (code=exited, status=1/FAILURE)
[rocky@cclr-dev-svn3-dr08-u14 ~]$ systemctl is-failed ansible-init
activating
Mine looks like that. Does it eventually give up?
There was a problem hiding this comment.
There was a problem hiding this comment.
What situation did this occur in? If its during dev with a broken compute-init I think I'm not too fussed and we should just close this PR - Bertie is going to add some notes to the docs on how to recover this situation. If it occurs in production its broken anyway, and ansible can't recover it really, IMO.
There was a problem hiding this comment.
Correct, this was a the issues adding new nodes into the cluster. I guess it should be resolved by Bertie's work on making it exit early if the hostvars do not exist.
For other issues though... It is a large timeout, so would be nice if it failed slightly quicker IMO.
There was a problem hiding this comment.
Yeah, maybe we should knock it down considerably, that's not this PR though. I'm going to close this and we can address during review of the compute-init update.
|
Closed on the basis this isn't going to help as ansible-init keeps retrying. Timeout in compute-init should be knocked down though if possible on #627. |
Avoids waiting for a timeout if ansible-init happens to have failed already before running site.