John Strunk

@jarrpa Could you please provide a project status in the README.md as OP requested? For Gluster in Kubernetes, folks probably want to look at https://github.com/kadalu/kadalu

First thing to check is your network bandwidth. Gluster replicates from the clients, so writing at 33MB/s will use 100MB/s of network BW ~1GbE. (Assuming 3 way replication) On Dec...

@nkiesel Are you still having unexpected performance or can this issue be closed?

For network throughput, I'm a bigger fan of ttcp: * Receiver: `ttcp -rs -n 102400` * Sender: `ttcp -ts -n 102400 dest_host` That'll send ~800 MB between the hosts and...

I'm assuming @martinsimon007 and @nkiesel work together? If not, lets split this into two different issues. Going on the assumption this is the same system... So we've got ~510 MB/s...

> I think we can provide an option in the configuration file Please make sure this is available via env or command line also. Managing config files is difficult in...

"tags" should be key=value pairs. While they aren't to be interpreted by "core", it's fair game for plugins and peripheral entities to use them. For example, the operator will communicate...

Thoughts on operator/gd2 synchronization are at: https://github.com/gluster/anthill/pull/36 I believe we can implement the boolean flags via node-level tags, so I don't think this is an API change per-se, just tags...

Excellent. Can you please put together a PR that documents how to deploy gluster-prometheus such that it will interact properly w/ GD2's leader election? It is still unclear to me...

Since this is credentials, I assume the json file would be in a Secret, not a PVC?