I've got a cluster of 3 CoreOS machines running on Azure. I rebooted two at the same time and the cluster failed as expected.

I need to replace the discovery token, cloud-config is read on every bootup, but according to the CoreOS docs:

Once an instance is provisioned on Azure, the cloud-config cannot be modified.

Is there a method to recover from this, short of destroying and then deploying a new cluster ?


there configuration file exists in location



sudo vim /var/lib/waagent/CustomData

You should be able to edit it. After the reboot configuration will pick up.


You could try modifying the etcd service definition in /run/systemd/system/etcd.service.d/20-cloudinit.conf - you should see something like


Modify the discovery URL to your new one, then restart it

systemctl daemon-reload
systemctl restart etcd

You will need to test if this survives a reboot on Azure though!

  • It doesn't survive a reboot. I might start looking for a new host. – Mark Jul 5 '15 at 22:30

If you remove two nodes in three node cluster, you loose the quorum, with 3 nodes you can only with lose one node, for more information about the fault tolerance for the CoreOS:

Fault Tolerance Table

It is recommended to have an odd number of members in a cluster. Having an odd cluster size doesn't change the number needed for majority, but you gain a higher tolerance for failure by adding the extra member. You can see this in practice when comparing even and odd sized clusters:
Cluster Size    Majority    Failure Tolerance
1   1   0
3   2   1
4   3   1
5   3   2
6   4   2
7   4   3
8   5   3
9   5   4


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.