1

Over the past updates (espcially the update to etcd-wrapper: https://github.com/coreos/coreos-overlay/tree/master/app-admin/etcd-wrapper) my etcd-cluster with five machines broke. Now I'm trying to recover from that.

Currently I only use it for the etcd-lock while updating the machines. So there is basically no important information in there (yet).

I tried a lot of stuff and since, in my opinion, the documentation is not that structured, I mixed a lot of infos form different parts together. Currently I'm in a state, where one node wants to elect a leader with another node. But that other node thinks it is part of another cluster. And I cannot do anything on the first one anymore. etcdctl just revokes all actions with the info, that the cluster is not healthy...

So I want to do a fresh start again!

Could someone point me to some part of the documentation, that explains that? It seems, that the disaster recovery part (https://coreos.com/etcd/docs/latest/op-guide/recovery.html), that allegedly contained the info, has changed (and is also depreacted).

Additionally it would be really helpful, if someone could point me to the part of the docu, that explains how to setup a etcd-cluster without re-installing the whole system.

Thanks in advance!

PS: I would have gladly selected some better tags (like coreos, etcd, etcd-cluster), but I am not allowed to create them now :)

0

I managed to fix it. First I disabled and deleted all etcd2-related units. Then I created a new dropin to the etcd-member.service with the following content:

[Service]
Environment="ETCD_OPTS=--name cluster00 \
  --discovery https://discovery.etcd.io/<MY_ID> \
  --listen-client-urls http://<NODE_IP>:2379,http://127.0.0.1:2379 \
  --listen-peer-urls http://<NODE_IP>:2380 \
  --advertise-client-urls http://<NODE_IP>:2379 \
  --initial-advertise-peer-urls http://<NODE_IP>:2380"

I generated a new discovery id for the cluster.

Then I teared down the old cluster completely, by deleting all content of /var/lib/etcd/* and the directory /var/lib/etcd2 completely and (re-)started the etcd-member service.

Be aware, that I don't have ssl enabled at this point!

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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