"norman.kern" <norman.kern@xxxxxxx> writes:
> If you want to use ceph , you have to know ceph+k8s+rook, each module
> can cause problems.
I think this is a very good point to consider. With the introduction of
cephadm, we actually had to choose:
We were coming from Devuan + Self written scripts + Ceph Nautilus.
The question for us was switching over to
a) a systemd based Linux distro and cephadm
b) or use k8s + any Linux distribution (in our case: Alpine Linux)
For us it it looked like cephadm is actually providing similar features
to k8s in regards to deployment. However with k8s as the underlay and
the experience with ceph-deploy and ceph-ansible, we felt that using a
standardised deployment (k8s) plus a ceph specific tooling (rook) is
long term more sustainable.
Obviously, only time will tell which is the smarter choice, however for
us it also comes with the advantage of easily deploying workload on top
of ceph and to create CephFS and RBD instances "on the fly".
That said, rook can be used to only supply a ceph cluster, but you get
much more out of the box in regards of "using the ceph cluster".
This might be something you want or to dislike.
Best,
Nico
--
Sustainable and modern Infrastructures by ungleich.ch
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
