Start one unit file automatically after another unit file - Not working

[Karthik Jayaraman]

abc-epg@.service Unit File

[Unit]

Description=abc EPG Service

Wants=epg-discovery@%i.service

Before=epg-discovery@%i.service

# Requirements

Requires=etcd.service

Requires=docker.service

# Dependency ordering

After=etcd.service

After=docker.service

[Service]

User=core

# Let processes take awhile to start up (for first run Docker containers)

TimeoutStartSec=0

# Change killmode from "control-group" to "none" to let Docker remove

# work correctly.

KillMode=none

# Get CoreOS environmental variables

EnvironmentFile=/etc/environment

# Pre-start and Start

## Directives with "=-" are allowed to fail without consequence

ExecStartPre=-/usr/bin/docker kill abc-epg

ExecStartPre=-/usr/bin/docker rm abc-epg

ExecStartPre=/usr/bin/docker -D  pull abcuser/abc-repository:abc-epg

ExecStart=/usr/bin/docker run --name abc-epg -p 8080:8080 abcuser/abc-repository:abc-epg

# Stop

ExecStop=/usr/bin/docker stop abc-epg

TimeoutStartSec=5m

TimeoutStopSec=20s

RestartSec=30s

[X-Fleet]

X-Conflicts=abc-epg@*.service

epg-discovery@.service Unit File

[Unit]

Description=abc etcd Service

# Requirements

Requires=etcd.service

# Dependency ordering

After=etcd.service

After=abc-epg@%i.service

BindsTo=abc-epg@%i.service

[Service]

User=core

# Let processes take awhile to start up (for first run Docker containers)

TimeoutStartSec=0

# Change killmode from "control-group" to "none" to let Docker remove

# work correctly.

KillMode=none

# Get CoreOS environmental variables

EnvironmentFile=/etc/environment

# Pre-start and Start

## Directives with "=-" are allowed to fail without consequence

ExecStart=/bin/bash -c '\

  while true; do \

    curl -f ${COREOS_PUBLIC_IPV4}:8080; \

    if [ $? -eq 0 ]; then \

      etcdctl set /services/epg/${COREOS_PUBLIC_IPV4} ${COREOS_PUBLIC_IPV4}:8080 --ttl 30; \

    else \

      etcdctl rm /services/epg/${COREOS_PUBLIC_IPV4}; \

    fi; \

    sleep 20; \

  done'

[X-Fleet]

X-ConditionMachineOf=abc-epg@%i.service

With the aforementioned configuration, with wants and after, the discovery unit must be started right after the service unit right ? But eventhough the service unit is started successfully, the discovery unit is not. Any problems with my config. ?

[Brian Waldon]

What fleetctl commands are you using? The Wants/Before options are only consumed by systemd, not by fleet, so you do have to make sure the units are scheduled (fleetctl load) before starting abc...@N.service.

--
You received this message because you are subscribed to the Google Groups "CoreOS User" group.
To unsubscribe from this group and stop receiving emails from it, send an email to coreos-user...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Karthik Jayaraman]

thanks for your prompt response. I just now found that. 

[Karthik Jayaraman]

thanks for your prompt response. I just now found that. 

On Friday, October 3, 2014 9:29:18 AM UTC-7, Brian Waldon wrote:

[Karthik Jayaraman]

Even after i load units, the loaded unit does not start by itself.

fleetctl submit tve-epg\@.service epg-discovery\@.service

fleetctl submit tve-epg\@{1..2}.service  epg-discovery\@{1..2}.service

fleetctl load tve-epg\@{1..2}.service  epg-discovery\@{1..2}.service

--------------------------------------------------------------------------------------

Unit tve...@1.service loaded on 0fc6daa6.../292.268.222.24

Unit tve...@2.service loaded on 275df756.../292.268.222.27

Unit epg-di...@1.service loaded on 0fc6daa6.../292.268.222.24

Unit epg-di...@2.service loaded on 275df756.../292.268.222.27

--------------------------------------------------------------------------------------

fleetctl start tve-epg\@{1..2}.service

--------------------------------------------------------------------------------------

Unit tve...@1.service launched

Unit tve...@2.service launched on 275df756.../192.168.111.17

--------------------------------------------------------------------------------------

fleetctl list-units

--------------------------------------------------------------------------------------

UNIT                    MACHINE                         ACTIVE          SUB

epg-di...@1.service 0fc6daa6.../192.168.111.14      inactive        dead

epg-di...@2.service 275df756.../192.168.111.17      inactive        dead

tve...@1.service       0fc6daa6.../192.168.111.14      active          running

tve...@2.service       275df756.../192.168.111.17      active          running

--------------------------------------------------------------------------------------

[Jonathan Boulle]

Can you file an issue against fleet to troubleshoot this further?

[Karthik Jayaraman]

Thanks for your response. The issue was with my unit files. Corrected them and loaded them as per my constraints. It worked. Thanks.