Open jmccoy555 opened 6 years ago
Progress.. updating user-data to;
coreos: etcd: advertise-client-urls: "http://${ETCD_IP_ADDRESS}:2379" initial-advertise-peer-urls: "http://${ETCD_IP_ADDRESS}:2380" listen-client-urls: "http://0.0.0.0:2379,http://0.0.0.0:4001" listen-peer-urls: "http://${ETCD_IP_ADDRESS}:2380,http://${ETCD_IP_ADDRESS}:7001" name: "${ETCD_HOSTNAME}" initial-cluster: "${ETCD_HOSTNAME}=http://${ETCD_IP_ADDRESS}:2380" update: reboot-strategy: "reboot" units:
- name: etcd-member.service command: start ................
Allows the service to start.
Now I'm just getting;
etcdctl member list Error: client: etcd cluster is unavailable or misconfigured; error #0: dial tcp 10.10.1.140:4001: connect: connection refused
from the worker.
Well, if I remember right, the last time I looked at this CoreOS had changed the way they configure tools like etcd on startup now. These scripts haven't really been changed to use that new format. I'm sure it could be done, but I no longer use etcd this way so I haven't bothered to do it myself.
If you come up with something that works it would sure be interesting to see.
It looks like you are quite right - clustering.
Also, fleet is no longer developed or included with coreOS, so theres a second problem, but it doesn't look like it would be too difficult if using static IPs.
I think I'm going to deploy with the script but then use Docker Swarm (for now anyway).
Right, fleet is basically dead - and rightfully so with the likes of kubernetes and swarm using a better, more supported model. But I still do use the script to deploy new standalone CoreOS servers
Hi,
Just trying this out (once I worked out I needed to remove the -s flag in the create-cluster script) everything appears to deploy nicely, however user-configdrive is failing;
The config appears to be loaded (I removed a few of the ssh keys to shorten the text);
Net interface checked;
EDIT:
Running on a standalone esxi 6.7
Any ideas or pointers would be appreciated as this really looks like a cool way to deploy.
Thanks.