Single machine / multiple monitors
Hi, I’m getting conflicting reads from the documentation. It seems that by using the “cluster name”[1], multiple clusters can be run in parallel on the same hardware. In trying to set this up with `ceph-deploy`, I see the man page[2] says "if it finds the distro.init to be sysvinit (Fedora, CentOS/RHEL etc), it doesn't allow installation with custom cluster name and uses the default name ceph for the cluster”. Is it possible to run multiple clusters on the same hardware with CentOS 7 as the base OS? Thanks, Brian [1] https://docs.ceph.com/docs/nautilus/install/manual-deployment/#monitor-boots... [2] https://docs.ceph.com/docs/nautilus/man/8/ceph-deploy/?highlight=ceph-deploy...
Custom cluster names are being incrementally deprecated. With ceph-deploy I thought they were removed in 1.39. You could probably achieve parallel clusters with containerized daemons if you tried hard enough, but I have to ask what leads you to want to do this.
On Mar 11, 2020, at 4:07 PM, Brian Topping <brian.topping@gmail.com> wrote:
Hi, I’m getting conflicting reads from the documentation. It seems that by using the “cluster name”[1], multiple clusters can be run in parallel on the same hardware.
In trying to set this up with `ceph-deploy`, I see the man page[2] says "if it finds the distro.init to be sysvinit (Fedora, CentOS/RHEL etc), it doesn't allow installation with custom cluster name and uses the default name ceph for the cluster”.
Is it possible to run multiple clusters on the same hardware with CentOS 7 as the base OS?
Thanks, Brian
[1] https://docs.ceph.com/docs/nautilus/install/manual-deployment/#monitor-boots... [2] https://docs.ceph.com/docs/nautilus/man/8/ceph-deploy/?highlight=ceph-deploy... _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
On Mar 11, 2020, at 5:28 PM, Anthony D'Atri <anthony.datri@gmail.com> wrote:
Custom cluster names are being incrementally deprecated. With ceph-deploy I thought they were removed in 1.39.
Maybe the man page is not up-to-date. The command does not accept `—cluster` as an option.
You could probably achieve parallel clusters with containerized daemons if you tried hard enough, but I have to ask what leads you to want to do this.
Yes, that’s an interesting thought. I noticed on Octopus man pages today that container support is coming and the repos already have containers with full version history for Nautilus. Use case is a deployment with limited connectivity between one of three nodes (~350Kbps / 5ms RT). I’m currently running production with a single node monitor, but am trying to get to a point where all three nodes have monitors. The limited connectivity node would *only* have a monitor, no OSDs, MDS, etc. This is all possible with a single cluster, but this limited node also needs storage. Maybe a hybrid solution would be for the limited node to run the monitor as above on bare metal and Rook running for local needs out of containers. I guess another option is to throw in the towel on bare metal storage and go full Rook in both locations
participants (2)
-
Anthony D'Atri
-
Brian Topping