I get the following when trying to update:
Jul 22 13:25:09 appliance zincati: [ERROR] failed to check Cincinnati for updates: client-side error: error sending request for url (https://updates.coreos.stg.fedoraproject.org/v1/graph?os_version=31.20200323.3.2&basearch=x86_64&platform=vmware&group=default&os_checksum=9b322b9cead87ebf80881b75560e481b6e5a26966814b01bf0c12b61fa363708&node_uuid=d41ec79bb77b41c9877462338f6cc10b&stream=stable): error trying to connect: tcp connect error: Connection timed out (os error 110)
Is the update server down?
Thanks for reporting. Someone from the Fedora CoreOS team is looking into it.
Sometimes we have infrastructure issues and the update server become unavailable. It’s back now and I just confirmed a node started on a previous version is able to update.
Still down for me, but I’ll keep an eye on it. If I still can’t connect to the update server over the next couple days, I’ll let you know just in case there is something else going on.
Ahh. I think I see your issue. It looks like your node is old? Our update server was migrated to the non-staging location and the URL was updated at the same time. The staging server stayed up and running for a while, but the recent Fedora Infrastructure datacenter migration caused it to go away and it’s no longer running. Most nodes should have updated by then, but it looks like yours didn’t. You’re running
31.20200323.3.2 (built on April 8th).
Have these nodes been powered down/disconnected for a long time or do you have automatic updates disabled?
The following should get you unstuck:
$ sudo tee /run/zincati/config.d/99-fixes.toml << EOF
allow_downgrade = true
$ sudo systemctl restart zincati.service
The system will progress through updates (including a Fedora 32 rebase) and reboot a few times.
Thanks @dustymabe! That was it. The FCOS hosts we deploy are often in environments which isolate hosts from the internet. It isn’t unusual for these hosts to go months without an OS update.