Consul service panics on journalctl restart

[xytis]

The same vulnerability as with docker: moby/moby#19728

consul version for Client

Client: Consul v0.6.3

consul info for Client

Client:

agent:
    check_monitors = 0
    check_ttls = 0
    checks = 1
    services = 1
build:
    prerelease =
    revision = c933efde
    version = 0.6.3
consul:
    known_servers = 5
    server = false
runtime:
    arch = amd64
    cpu_count = 24
    goroutines = 48
    max_procs = 2
    os = linux
    version = go1.5.3
serf_lan:
    encrypted = false
    event_queue = 0
    event_time = 62
    failed = 0
    intent_queue = 0
    left = 0
    member_time = 601
    members = 34
    query_queue = 0
    query_time = 1

Description of the Issue (and unexpected/desired result)

consul daemon restarts when journald is restarted

Reproduction steps

systemctl restart systemd-journald.service

[rbjorklin]

We've also encountered this problem. Not sure if it matters but we've been running consul as a service too.

[Unit]
Description=replicated key—value store
After=network-online.target
[Service]
User=consul
ExecStart=/usr/local/bin/consul agent -config-file /etc/consul.d/config.json
KillSignal=SIGINT
[Install]
WantedBy=default.target

EDIT:

# /usr/local/bin/consul version
Consul v0.6.4
Consul Protocol: 3 (Understands back to: 1)

[rbjorklin]

Reading through the thread over at the Docker project this issue might have been resolved in Consul >= 0.7.0 because Consul was built with golang >= 1.6. #1963

Will try this out and report back.

[xytis]

In addition to golang update such line has to exist:
hashicorp/nomad#1798
(hashicorp/nomad@3a4bb56)

[mckennajones]

Was unable to reproduce the issue on consul version 0.7.0. However, it still seems like a good idea to add the code that ignores sigpipe signals, so I created a quick PR.