Node does not start automatically after reboot.

[bipsendk]

For other reasons, I had to reboot my test server with the single node installation.
After reboot, the single-node ECS installation does not start the services.

What should be done in order to start everything up again - in order to avoid having to start an installation from scratch?

[padthaitofuhot]

Thanks for filing an issue on this.

Disposition

Enabling the ECS system to boot back up after a system reboot should be included in the install script. This enhancement is marked for inclusion in the 2.1 release due out sometime next quarter.

Work-Around

1. Enable the docker service to run on boot.
   systemctl enable docker.service
2. Start the docker container:
   docker start <Your container ID here>
3. Add a line to /etc/rc.local to start the docker container on reboot:
   echo "docker start <Your container ID here>" >>/etc/rc.local
4. Update /etc/fstab*** to make sure the data disk gets mounted on reboot. For example, assuming one data disk as sdb, add:
   /dev/sdb1 /ecs/uuid-1 xfs defaults 0 0
5. Then run mount -a to mount all the block devices.
6. Increase the memory limit for pids by running:
   sysctl -w kernel.pid_max=655360
7. Add that same line (without sysctl -w) to /etc/sysctl.conf so it's enabled on reboot:
   echo "kernel.pid_max=655360" >>/etc/sysctl.conf

*** For reference, here's an excerpt of mount output on one of my single-node deployments with five block devices in the datastore:

/dev/sdc1 on /ecs/uuid-1 type xfs (rw,noatime,seclabel,attr2,inode64,noquota)
/dev/sdd1 on /ecs/uuid-2 type xfs (rw,noatime,seclabel,attr2,inode64,noquota)
/dev/sde1 on /ecs/uuid-3 type xfs (rw,noatime,seclabel,attr2,inode64,noquota)
/dev/sdf1 on /ecs/uuid-4 type xfs (rw,noatime,seclabel,attr2,inode64,noquota)
/dev/sdg1 on /ecs/uuid-5 type xfs (rw,noatime,seclabel,attr2,inode64,noquota)

[bipsendk]

What is the container ID (where do I find it)? It does not seem to be a parameter specified during installation/configuration.

[Callero]

Please issue sudo docker ps and you should be able to see a container ID

[bipsendk]

docker ps

FATA[0000] Get http:///var/run/docker.sock/v1.16/containers/json: dial unix /var/run/docker.sock: no such file or directory. Are you trying to connect to a TLS-enabled daemon without TLS?

[bipsendk]

After
systemctl enable docker.service
there should be a
systemctl start docker.service
and then (for standalone):
docker start ecsstandalone

[Callero]

Please remember to use sudo

[bipsendk]

I am logged in as root :-)

Status so far 5 minutes after reboot and startup:

tcp 0 0 0.0.0.0:9099 0.0.0.0:* LISTEN 22724/storageserver
tcp 0 0 0.0.0.0:9069 0.0.0.0:* LISTEN 22724/storageserver
tcp 0 0 10.11.12.152:9204 0.0.0.0:* LISTEN 22704/vnest
tcp 0 0 0.0.0.0:9205 0.0.0.0:* LISTEN 22704/vnest
tcp 0 0 0.0.0.0:9206 0.0.0.0:* LISTEN 22704/vnest
tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 1045/sshd
tcp 0 0 0.0.0.0:64443 0.0.0.0:* LISTEN 22796/ecsportalsvc
tcp6 0 0 :::22 :::* LISTEN 1045/sshd

5 minutes later, the system is up and running

[jasoncwik]

Depends on issue #66. Also need to install systemd scripts to ensure docker is auto-started and the docker image is auto-started after docker.

[keithxanderson]

Hello, I've followed the single node software only install process for ECS 2.2 plus the workarounds listed above but I can't log into the webUI after a reboot. Any tips in how to get this working? I log in as root with the correct password and the page hangs with a spinning "login" animation. Have CentOS 7 running in a VM.

[APeter0807]

Resolved the above issue in discussion elsewhere.

When editing /etc/fstab, use the following settings rather than the default:

/dev/sdb1 /ecs/uuid-1 xfs rw,noatime,seclabel,attr2,inode64,noquota 0 0

[victorock]

Please see pull request #108

[bottkars]

/etc/rc.local is a symlink to /etc/rc.d/rc.local
even if you put stuff into /etc/rc.local, you mast chmod the symlink in /etc/rc.d/rc.local

so chmod +x /etc/rc.d/rc.local should work !

[padthaitofuhot]

This is working in installer 2.0.0; unfortunately release has been delayed.