Howdy,
I am in the process of moving from Zimbra OSE to Carbonio CE, and I noticed that after completing a single server install I am unable to start the 'carbonio-prometheus' service. It cites a permissions issue on /etc/carbonio/carbonio-prometheus/prometheus.yml.
I'm running Rocky 8, so I am unsure if this permissions issue is a problem with the RPM package not installing with the correct permissions or not..
Fixing the permissions issue by changing the group ownership of the affected file to the carbonio-prometheus group fixes the startup issue, but now I am seeing a repeated error in the log:
ts=2023-10-30T20:53:15.336Z caller=consul.go:293 level=error component="discovery manager scrape" discovery=consul msg="Error retrieving datacenter name" err="Unexpected response code: 403 (Permission denied)"
I think(?) this is due to an issue with the ACL policy that is being configured for consul, since prometheus can connect to whatever service it's supposed to it's just being denied access to the information it is requesting. Is this intended behaviour? Is this service supposed to be used on a single server installation?
Whereabouts is the right place to open a bug report if this is one?
Thanks in advance 🙂
This issue is still prevalent in the latest release of Carbonio CE (23.11.0)
I am also seeing a missing configuration file for carbonio-prometheus-mysqld-exporter now, which causes that service to fail to start as well.
See my log snapshot from the systemd service:
The file does not appear to exist at all.
[zextras@hermes carbonio-prometheus-mysqld-exporter]$ pwd /etc/carbonio/carbonio-prometheus-mysqld-exporter [zextras@hermes carbonio-prometheus-mysqld-exporter]$ ls -l total 0
Hi,
Sorry to hear about the trouble.
Right now Ubuntu 20.04.6 LTS and RHEL8 are the only officially supported OS for Carbonio CE. Therefore, we would request to all of our users through you that please use either one of these distros.
Developing the system (Carbonio CE) for these two supported OS takes lots of detailed and complex plan and process and even after that we get bugs in these platforms after a new release. So diagnosing an issue on a different platform (other than the supported OS) hinders the development plan, schedule and process for each month.
In future considering our users preferences, we may include more OSs in the supported OS list but till then our humble request would be to use the supported OS only.
We hope you could understand the situation.
Thanks and regards,
Sharif
This is happening to me too on a fresh install of Ubuntu 20.04. It was working on an installation a few months ago so there's some regression in new versions. I'm using officially supported distro and follow the single-server installation instructions down to the letter.
Here's another report that's not resolved either, dating to September: https://community.zextras.com/forum/carbonio-setup/the-issues-while-installing-follow-the-oficial-guide/
So, I found the "fix": https://community.zextras.com/forum/carbonio-setup/installation-and-upgradation-of-carbonio-ce/#post-4904
The documentation is wrong, bootstrap should be done before restarting the service. This was reported in April.
The installation procedures are routinely carried out by our QA at every release following the installation on docs. Can you provide more information about the problem? Which documentation is wrong (community or docs) and where?
@stefanodavid In single-server installation of community edition (link), step 3 says to restart carbonio-prometheus-nginx-exporter.service, and then proceed to firewall configuration and carbonio-bootstrap in step 5. This is wrong because carbonio-prometheus-nginx-exporter will keep failing until carbonio-bootstrap is run. Only then will the carbonio-prometheus-nginx-exporter service start properly.
@danijel-tudek Thanks for reporting, I guess I misplaced it. I am currentl completely reworking the installation, so I'll keep your comment present