-
Notifications
You must be signed in to change notification settings - Fork 47
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add glusterfs as storage #17
Comments
While I don't have experience using GlusterFS with Docker. looking at the current state of the Docker volume plugin glusterfs-volume-plugin it appears that it's a manual operation to create GlusterFS volumes outside of Docker Compose and then reference those existing GlusterFS volumes from within compose files. By comparison, the Portworx Docker volume plugin will automatically create non-existing volumes whenever a stack is (re)deployed. There seem to be ansible playbooks out there that can help one deploy GlusterFS clusters across hosts. By possibly using such a playbook to set up a GlusterFS cluster and then manually creating the appropriate volumes, it should then be possible to replace any "driver: pwx" (Portworx) volumes defined in swarmstack's docker-compose.yml file with that volume plugin equivalents, for https://github.com/trajano/docker-volume-plugins/tree/master/glusterfs-volume-plugin this looks to be "driver, driver_opts, name". A similar technique would probably cover the Docker Compose changes needed to support other HA storage options out there, at least for those for which a Docker volume plugin already exists. If a swarmstack user experienced with GlusterFS and Docker wants to research and write up a specific procedure for deploying GlusterFS and bringing up swarmstack (using an ansible playbook to deploy the GlusterFS cluster), I'd accept a PR for the documentation and would also look into generalizing or integrating the storage definitions within swarmstack's ansible/cluster YAML file so that it might potentially remain the single file requiring user editing before executing a (Portworx, GlusterFS, etc) storage playbook. I've no problem forking such an ansible GlusterFS playbook and adding the necessary changes so that it could reference that anslble/swarmstack cluster definition. While I don't have the time at present myself to research and test GlusterFS with Docker, I'd look to someone experienced with the above technologies to confirm that it's solid and reliable (volume mounts don't get stuck on a specific host when the containers move) when used with Docker for production workloads, this has been my experience with Portworx even using their free developer version. |
Hey Hello, Im interested on this too, because portworx has a 5 node and 5TB limit, and rexray its unmaintained, theres an issue talking about it rexray/rexray#1338 I hope to help on any of those projects, or even an swarm adaptation for openEBS, I dont know which path would be easier and I need to research more |
This looks really great, one thing that would make it even better was to use opensource storage solutions like glusterFS, it would be small changes and make this project relay on all open software
The text was updated successfully, but these errors were encountered: