[Gluster-users] Production cluster planning

[Gandalf Corvotempesta]

2016-10-27 0:14 GMT+02:00 Joe Julian <joe at julianfamily.org>:
> For now I can say that gluster performs better and has a much better
> worst-case resolution. If everything else goes to hell, I have disks with
> files on them that I can recover on a laptop if I have to.

Totally agree.


> Of course when you ask the Inktank consultants (now Red Hat) about "What
> happens when it fails?" the answer is "It doesn't fail." Well guess what...

It doesn't fail?
There was a recent HUGE outage in DreamHost cluster, some days ago. It
took (if i remember properly 2 days to fix, and this was dreamhost,
not a "normal" company)
There was another huge outage some weeks ago, as seen on ceph ML.
Is not the first time that I see ceph clusters totally down or with
huge corruption (LevelDB corrupted and so on)

So, it fails, like any other software. As I always say, having the
file untouched and easily recoverable (without shards, you can simply
run "rsync", with shard you run rsync+some script to mege each chunk)
is the real value added, when everything is gonna bad.