0

I have a server which content is on a gluster mount. The underlying brick is on the same machine.

I was wondering if there is a possibility to set up some kind of fallback in case gluster crashes/disconnects/goes on vacations.

What I have in mind is something like mounting the brick as a bind mount first, then gluster on the same mountpoint, thus masking the first mount, but if the gluster mount is umounted, the server access the files in the brick through the bind mount.

The point being that the server content still remains accessible to the server software.

Any thoughts on this?

greg

Improve this question

1 Answer 1

Reset to default
0

This is not recommended as accessing the underlying filesystem without using gluster would corrupt the volume.

You are using gluster to make a service HA via another server so I would shut down all services on this node and let the other one take over in case the gluster volume is not mounted.

Improve this answer
2
  • Thanks for your answer. My preocuppation is that - in the recommended setup - if the gluster daemon crashes, the files are inaccessible making the web server "bare"
    – greg
    Apr 22, 2014 at 23:47
  • Well yes, but then the service should stop and the other server should take over. I don't know what services you are running and how, but if you've got for example a virtual IP with keepalived then you can add a check for gluster that will lower the priority when it's not accessible.
    – Jure1873
    Apr 23, 2014 at 18:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .