Volumes are convenient and persistence is a blessing ... except when they're not. We moved to using Docker swarm the other day, and realized that vanilla volumes are not replicated across a swarm. Rather, if a service changes its node and ends up on a different host, we also end up with a new volume (or an outdated one from the last time the container instance lit up on that node). Yes, with volumes you can use different driver types, like NFS (or Glusterfs if you want to be fancy) ... but it gets messy. So we ended up just using fixed database services -the main thing for us requiring persistent volumes- external to our swarm.
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
Volumes are convenient and persistence is a blessing ... except when they're not. We moved to using Docker swarm the other day, and realized that vanilla volumes are not replicated across a swarm. Rather, if a service changes its node and ends up on a different host, we also end up with a new volume (or an outdated one from the last time the container instance lit up on that node). Yes, with volumes you can use different driver types, like NFS (or Glusterfs if you want to be fancy) ... but it gets messy. So we ended up just using fixed database services -the main thing for us requiring persistent volumes- external to our swarm.