I'm looking to give glusterfs a go and have a few questions that I haven't found definitely answers to. I was wondering if I might be able to poll those that have gone down this path already :) Sorry for the barrage of questions, these are the remaining questions that I need answers to green-light the project.
1. currently using cachefilesd on my clients with nfsv4 with a decent performance gain. Will cachefilesd work with a nfs-compliant gluster mount? Does the native FUSE client offer similar local caching capability?
2. I'm most interested in the distributed mode of operating bricks, as I'll still provide resiliency via hardware RAID6. My plan is to get a single brick up and running and then add a second one in distributed mode later. The likely scenario is this first brick will be at 80% capacity when I add the second. What happens in this case, are the files balanced between the two when the second brick is added?
3. I have an existing file system, with permissions, etc. on the machine already that I'm about to setup glusterfs on. Can I simply use this directory structure when creating the volumes, thus bringing up my first instance with all files intact and ready to go? Will this still allow use of things like libgfapi? Is there an example somewhere showing the commands to bring and existing filesystem online as a glusterfs?
4. In the above scenario, will all existing permissions/ownership/etc. be retained?
5. In a cluster consisting solely of distributed bricks, is it okay to also mount the filesystem locally to each on of these bricks? I have some lightweight jobs that will be adding files to the filesystem and would like to run these directly from the distributed bricks, if possible.
Thanks!!
_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-users
