Description of problem: From time to time, my Amazon or Google cloud providers disconnect. This causes the providers to fail to authenticate and set their status to 'Validation Required'. From there on, it is impossible to reconnect the providers without modifying their value (e.g. name) first, by re-validating the credentials from the edit screen and modifying their properties to be able to save the valid state. This is an issue as I would expect the providers to be able to reconnect on their own. Version-Release number of selected component (if applicable): 5.6.1.2.20160810181333_8ba817b How reproducible: Always Steps to Reproduce: This is hard to reproduce, AWS or Google fail from time to time (e.g. network) 1. From Compute > Clouds > Providers, try to refresh a provider with a refresh status of 'Failure' Actual results: Refresh does not work as 'Credentials' are not validated. Expected results: Credentials haven't changed and are still valid. Additional info: The only way to reconnect is to validate the credentials again from its edit screen and change a value in order to save it. Then Refresh and re-check authentication will work.
https://github.com/ManageIQ/manageiq/pull/11297
For the UI, you can use the "Authentication -> Re-check authentication status" dropdown. But saving a manager should also work. See the mentioned PR
In case the credential become marked invalid there is no way but through the UI to mark them valid again. This is because a refresh worker with it's credential marked invalid refuses to start. And there is no other process that I am aware of that checks credentials on a regular basis.
New commit detected on ManageIQ/manageiq/master: https://github.com/ManageIQ/manageiq/commit/b5021fde478b8bc1e5e89c2b8f8689a46007a842 commit b5021fde478b8bc1e5e89c2b8f8689a46007a842 Author: Aparna Karve <akarve> AuthorDate: Thu Sep 15 10:59:33 2016 -0700 Commit: Aparna Karve <akarve> CommitDate: Thu Sep 15 11:05:51 2016 -0700 Use $setViewValue to update the auth status returned from validate This sets the form to $dirty and hence saveable if validation stat changes from invalid to valid https://bugzilla.redhat.com/show_bug.cgi?id=1374865 .../ems_common/ems_common_form_controller.js | 32 ++++------------------ 1 file changed, 6 insertions(+), 26 deletions(-)
So, I checked by setting :authentication_check_interval: 1.minute in advanced settings and setting the credentials to become invalid by adding ec2.eu-central-1.amazonaws.com to /etc/hosts to point to localhost. The managers auth would become invalid, the refresher would stop to work. After correcting the hosts settings the schedule worker did an auth check and the refresh worker would start automatically again. So my Comment 6 https://bugzilla.redhat.com/show_bug.cgi?id=1374865#c6 is invalid. Changing the the issue to reflect the UI part only
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days