Closed mjura closed 5 years ago
And on even slower we just double that! (SCNR)
And on even slower we just double that! (SCNR)
@isbm nope, we don't need that, only default timeout value is too short
I think we néed to find a more convenient and granular solution instead of increasing timeout
Proposed solution is sufficient, default timeout value is too low. I was trying also implement this in following way:
ca-setup:
salt.state:
- tgt: 'roles:ca'
- tgt_type: grain
- tag: ca/setup_event
- highstate: True
wait-for-ca-setup:
salt.wait_for_event:
- name: ca/setup_event
- timeout: 120
- id_list:
- ca
generate-sa-key:
salt.state:
- tgt: 'roles:ca'
- tgt_type: grain
- tag: ca/generate_sa_event
- sls:
- kubernetes-common.generate-serviceaccount-key
wait-for-generate-sa:
salt.wait_for_event:
- name: ca/generate_sa_event
- timeout: 300
- id_list:
- ca
but it doesn't work as I expected.
@isbm Do you have maybe other proposal ?
After discussion with @ereslibre we agreed that default timeout 30 seconds can be littlebit too low for virtual environments. We would like increase it to 120 seconds.
On realy slow deployments salt-minion-ca can exceed timeouts for "ca-setup" and "generate-sa-key" modules. We can avoid this and add timeout option to "generate-sa-key".
Option "timeout" is the maximum time in seconds to wait before failing.