Scale GlusterFS Service

Hi,

I’m using the GlusterFS stack from the Rancher catalog. I did some experiments and it works fine, but I need some help scaling the service:

scaling the service (up and down) the gluster volume is not automatically fixed. With “gluster peer status” I see both new and old (disconnected) peers. With “gluster volume info” I see the original bricks configuration. Could you provide a guide to fix the glusterfs volume in case of service scaling? I need to replace bricks and peers. I took a look to the glusterfs documentation, but it’s not clearly explained…

We are just starting to add a more thorough readme for the GlusterFS service, but at this time, when scaling the GlusterFS service, the gluster volume will not be updated with the new nodes.

Thank you Denise,
I’ll do some experiments, if I find a solution I’ll post it there.

The intent of the catalog is to help have people get GlusterFS setup, and at this point in time, we don’t plan on providing a managed gluster solution.

@cloudnautique is working on improving our GlusterFS solution though. The difficult part is automatic management when new nodes are added.

1 Like

Perhaps using something like consul event handlers to manage the service?

well, you have built the GlusterFS in official catalog, and @deved is having difiiculties with a native function of gluster: isn’t it suppose to be a distributed FS? if when scaling the main feature of gluster is not working…is not “only” our problem…

also, “at this point” we are not asking for a managed gluster solution from a open source pioneer like you, we are asking support for a native function that actualy as per @deved issue, is not working.

any, and i mean ANY help is apreciated.

thank you