etcd service doesn't start on secondary nodes.
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Etcd Charm |
New
|
Undecided
|
Unassigned |
Bug Description
Opened by Pekkari on 2018-11-14 11:15:34+00:00 at https:/
-------
I'm currently trying to deploy an etcd cluster with juju, three nodes gets
installed and one become active idle, rest stuck at requesting certificates
to an easyrsa related:
Unit Workload Agent Machine Public address Ports Message │······
etcd/0* active idle 8/lxd/1 x.x.x.x 2379/tcp Errored with 0 known peers │······
etcd/1 maintenance idle 18/lxd/1 x.x.x.x Requesting tls certificates. │······
etcd/2 maintenance idle 28/lxd/1 x.x.x.x Requesting tls certificates.
Inside the broken nodes I try to start etcd service installed from snap and it never
succeed. Connectivity between nodes is ok though:
# systemctl status snap.etcd.etcd │······
● snap.etcd.
Loaded: loaded (/etc/systemd/
Active: failed (Result: start-limit-hit) since Wed 2018-11-14 09:37:59 UTC; 1h 35min ago │······
Process: 409798 ExecStart=
Main PID: 409798 (code=exited, status=1/FAILURE) │······
Nov 14 09:38:01 juju-d67a39-
Nov 14 09:38:01 juju-d67a39-
Nov 14 09:38:03 juju-d67a39-
Nov 14 09:38:03 juju-d67a39-
Nov 14 09:38:03 juju-d67a39-
Nov 14 09:38:03 juju-d67a39-
Nov 14 09:38:05 juju-d67a39-
Nov 14 09:38:05 juju-d67a39-
Nov 14 09:38:05 juju-d67a39-
Nov 14 09:38:05 juju-d67a39-
# systemctl start snap.etcd.etcd
# systemctl status snap.etcd.etcd │······
● snap.etcd.
Loaded: loaded (/etc/systemd/
Active: failed (Result: start-limit-hit) since Wed 2018-11-14 11:13:56 UTC; 50s ago │······
Process: 431109 ExecStart=
Main PID: 431109 (code=exited, status=1/FAILURE) │······
Nov 14 11:13:55 juju-d67a39-
Nov 14 11:13:55 juju-d67a39-
Nov 14 11:13:56 juju-d67a39-
Nov 14 11:13:56 juju-d67a39-
Nov 14 11:13:56 juju-d67a39-
Nov 14 11:13:56 juju-d67a39-
Nov 14 11:13:56 juju-d67a39-
Nov 14 11:13:56 juju-d67a39-
Thanks!
José.
=======
Comment created by tvansteenburgh on 2018-11-14 13:01:05+00:00
@Pekkari Please provide a bundle file so we can attempt to reproduce.
-------
Comment created by Cynerva on 2018-11-14 14:28:52+00:00
@Pekkari Can you check the status of the snap.etcd.etcd service on unit etcd/0? I suspect it's not running and you need to restart it.
The other units are likely failing because they can't contact the leader.
-------
Comment created by Pekkari on 2018-11-14 14:53:29+00:00
It was, charm was in active/idle status reporting Errored on 0 peers on the status line. The environment
was thrown out for now, I'll give proper output when I can reproduce it with a shorter bundle.