Branch: refs/heads/master
Home:
https://github.com/Checkmk/checkmk
Commit: fd8f925629cf8005480905b7f8a9268f54f72b7a
https://github.com/Checkmk/checkmk/commit/fd8f925629cf8005480905b7f8a9268f5…
Author: Giordano Tomassorri <giordano.tomassorri(a)checkmk.com>
Date: 2024-01-03 (Wed, 03 Jan 2024)
Changed paths:
A .werks/14943.md
Log Message:
-----------
14943 FIX Agent updater: allow automation user to login with password
Up to now, when using the agent updater CLI, you would have to use the
`--password` (or `-P`) parameter to specify the password for a human
user and the `--secret` (or `-S`) parameter to specify the secret for
an automation user.
The main reason for that is that the authentication flow is different
for the 2 cases.
This starts to be confusing with the interactive mode: in that case the
program assumes that you are using a human user and will fail if the
credentials that you enter are valid credentials for an automation user.
On top of that, the error message is completely misleading.
This gets even worse in the CSE, where only automation users can be used
for the registration, making it impossible to have this working with the
interactive mode.
With this commit, we are changing the behavior of the agent updater so
that an automation user credentials will work even if the secret is
specified with the `--password` (or `-P`) param.
This way the end user don't have to care about which param name is the
right one to use: they can just specify the password, or the secret,
with the `--password` param and it will work.
This also allows the interactive mode to work with an automation user.
This change is backward compatible, meaning that everything that used to
work up until now, will keep working even after this.
SAASDEV-944
Change-Id: I00403dbceab1b638ff4b179deeb2728ef75fab6c