What is etcd looking for in 127.0.0.1:4001? - linux

Question

I'm trying to set up a test cluster using etcd 2.3.7 installed from CentOS RPM on CentOS 7.1. On the Loader 1 I executed:
etcdctl member add loader2 http://10.11.51.231:2380
And received response which confirmed the operation completed successfully.
Similarly:
etcdctl member add loader3 http://10.11.51.231:2380
with all default settings, and here's what I see:
Loader 1 10.11.51.166
systemctl status etcd -ln1
etcd.service - Etcd Server
Loaded: loaded (/usr/lib/systemd/system/etcd.service; disabled)
Active: active (running) since Sun 2017-02-19 14:33:18 IST; 28min ago
Main PID: 19009 (etcd)
CGroup: /system.slice/etcd.service
└─19009 /usr/bin/etcd --name=default --data-dir=/var/lib/etcd/default.etcd --listen-client-urls=http://localhost:2379
Feb 19 15:02:03 loader3 etcd[19009]: cannot get the version of member a4803061db803edc (Get http://10.11.51.166:2380/version: dial tcp 10.11.51.166:2380: getsockopt: connection refused)
Tried to see cluster health:
etcdctl --debug cluster-health
Cluster-Endpoints: http://127.0.0.1:4001, http://127.0.0.1:2379
cURL Command: curl -X GET http://127.0.0.1:4001/v2/members
cURL Command: curl -X GET http://127.0.0.1:2379/v2/members
member ce2a822cea30bfca is unhealthy: got unhealthy result from http://localhost:2379
member da05b63349d818dc is unreachable: no available published client urls
cluster is unhealthy
Note how this ignores the two nodes added previously, but sends requests to random port on localhost...
Loader 2 10.11.51.174
At first this machine started OK, but after I saw there was something wrong with Loader 1, I tried adding Loader 1 as a member from this machine, and now I see the same picture on this machine too. I.e. it tries to query this 4001 port, where nobody responds. On all machines:
netstat -tupln | grep etcd
tcp 0 0 127.0.0.1:7001 0.0.0.0:* LISTEN 4507/etcd
tcp 0 0 127.0.0.1:2379 0.0.0.0:* LISTEN 4507/etcd
tcp 0 0 127.0.0.1:2380 0.0.0.0:* LISTEN 4507/etcd
Nobody listens on 4001...
Loader 3 10.11.51.231
On this loader I didn't try to add new members. So it looks like this:
etcdctl --debug cluster-health
Cluster-Endpoints: http://127.0.0.1:4001, http://127.0.0.1:2379
cURL Command: curl -X GET http://127.0.0.1:4001/v2/members
cURL Command: curl -X GET http://127.0.0.1:2379/v2/members
member ce2a822cea30bfca is healthy: got healthy result from http://localhost:2379
cluster is healthy
In other words it still sends requests to random port, but this time it isn't bothered by the fact that nobody replied...
Below is the contents of the configuration files:
cat /usr/lib/systemd/system/etcd.service
[Unit]
Description=Etcd Server
After=network.target
After=network-online.target
Wants=network-online.target
[Service]
Type=notify
WorkingDirectory=/var/lib/etcd/
EnvironmentFile=-/etc/etcd/etcd.conf
User=etcd
# set GOMAXPROCS to number of processors
ExecStart=/bin/bash -c "GOMAXPROCS=$(nproc) /usr/bin/etcd --name=\"${ETCD_NAME}\" --data-dir=\"${ETCD_DATA_DIR}\" --listen-client-urls=\"${ETCD_LISTEN_CLIENT_URLS}\""
Restart=on-failure
LimitNOFILE=65536
[Install]
WantedBy=multi-user.target
And:
cat /etc/etcd/etcd.conf
# [member]
ETCD_NAME=default
ETCD_DATA_DIR="/var/lib/etcd/default.etcd"
#ETCD_WAL_DIR=""
#ETCD_SNAPSHOT_COUNT="10000"
#ETCD_HEARTBEAT_INTERVAL="100"
#ETCD_ELECTION_TIMEOUT="1000"
#ETCD_LISTEN_PEER_URLS="http://localhost:2380"
ETCD_LISTEN_CLIENT_URLS="http://localhost:2379"
#ETCD_MAX_SNAPSHOTS="5"
#ETCD_MAX_WALS="5"
#ETCD_CORS=""
#
#[cluster]
#ETCD_INITIAL_ADVERTISE_PEER_URLS="http://localhost:2380"
# if you use different ETCD_NAME (e.g. test), set ETCD_INITIAL_CLUSTER value for this name, i.e. "test=http://..."
#ETCD_INITIAL_CLUSTER="default=http://localhost:2380"
#ETCD_INITIAL_CLUSTER_STATE="new"
#ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster"
ETCD_ADVERTISE_CLIENT_URLS="http://localhost:2379"
#ETCD_DISCOVERY=""
#ETCD_DISCOVERY_SRV=""
#ETCD_DISCOVERY_FALLBACK="proxy"
#ETCD_DISCOVERY_PROXY=""
#ETCD_STRICT_RECONFIG_CHECK="false"
#
#[proxy]
#ETCD_PROXY="off"
#ETCD_PROXY_FAILURE_WAIT="5000"
#ETCD_PROXY_REFRESH_INTERVAL="30000"
#ETCD_PROXY_DIAL_TIMEOUT="1000"
#ETCD_PROXY_WRITE_TIMEOUT="5000"
#ETCD_PROXY_READ_TIMEOUT="0"
#
#[security]
#ETCD_CERT_FILE=""
#ETCD_KEY_FILE=""
#ETCD_CLIENT_CERT_AUTH="false"
#ETCD_TRUSTED_CA_FILE=""
#ETCD_PEER_CERT_FILE=""
#ETCD_PEER_KEY_FILE=""
#ETCD_PEER_CLIENT_CERT_AUTH="false"
#ETCD_PEER_TRUSTED_CA_FILE=""
#
#[logging]
#ETCD_DEBUG="false"
# examples for -log-package-levels etcdserver=WARNING,security=DEBUG
#ETCD_LOG_PACKAGE_LEVELS=""
#
#[profiling]
#ETCD_ENABLE_PPROF="false"
So... what is going on? The error messages given by etcd are the typical mindless nonsense produced by Go built-ins. The HTTP server that etcd uses is again, the Go built-in junk, that produces non-standard and absolutely worthless replies. So I cannot understand what was (if at all) misconfigured / missing.