Please take a look at the following:
# pmm-admin --version
1.17.4
# pmm-admin config --server ...
Error restarting one of the services: "systemctl" failed: exit status 1
# systemctl list-units --failed
UNIT LOAD ACTIVE SUB DESCRIPTION
0 loaded units listed.
Some notes:
I already tried the --debug
- parameter and I confirm all calls are answered with HTTP 200
- There is no failed service
- As much as I love random guessing, seeing the command that failed with status 1 would be useful :)
I could raise a bug if there is a bug here (other than the unhelpful error) but first I’d appreciate some guidance in finding out what is going wrong.
Hi!
First, a dumb question, just in case 
Are you trying with root? Your prompt (#) suggests you are, but just in case…
Secondly, did you check https://jira.percona.com/browse/PMM-2123 ? SELinux, maybe? What OS and systemd version are you using?
Please send exact commands being used, and I would go ahead and submit a bug so that we add more information to that error message in any case.
Hi Agustin,
Yes I was root.
That bug seems a bit different to me.
No SELinux. And the problem persisted after disabling apparmor.
The system was Ubuntu 20.04 on EC2. Note that I didn’t test on a fresh machine htough.
I randomly discovered that runningpmm-admin repair twice solved the problem. These things happened consistently, even after removing and re-installing PMM Client.
It remains a total mystery for me, but I didn’t have more problems so I’m not spending more time on it.