Search results

cloud-init configures an account and the network settings inside a newly cloned VM. What is the connection to CephFS or RBD that is not working for you?

Yes, because the assumed failure zone is the host. If just an OSD fails it should be replaced. In small clusters the time to replace a failed disk is more crucial than in larger clusters where the data is more easily replicated to the other OSDs in the nodes.

There is a neat calculator at https://florian.ca/ceph-calculator/ that will show you how the set the nearfull ratio for a specific number of disks and nodes.

Check all Interfaces if the MTU has been set to 1500 again. Do not mix MTUs in the same Ethernet segment.

All these applications already replicate their data in the application level. They do not need a storage system that does this. Let these VMs run on local storage and you will get way better performance than Ceph with size=1.

https://rook.io/ can also be used to integrate an external Ceph cluster into Kubernetes.

IMHO this question should be asked on the Ceph mailing list.

If you keep the smaller disks the larger ones will get 8 times the IOPS because the size is one of the factors of Ceph's algorithms to distribute the data. The NVMe disks may be able to handle that.

Usually Ceph will move the data for you. Before you bring down a host drain and remove its OSDs. After swapping the disks create a new OSD on the new NVMe and Ceph will happily use it. After doing this with all nodes all your data will be on the new disks.

This is why I recommend a minimum of 5 nodes for Ceph.

Just the rule does not matter if there is no pool using the rule. You could re-configure all your pools to use the first rule "replicated_rule" when you only have the same type of SSDs in the cluster, but it may cause some backfilling traffic. After that you could remove both rules that...

You would need to have 5 nodes with Ceph OSDs to sustain 2 failed nodes. And even then they should not fail at the same time.

Das klingt aber nach einem Bug, oder? Denn warum sollte in dem Feld nicht der aktuelle Wert angezeigt werden?

Even with a cephadm orchestrated cluster the NVMe-oF feature is currently not production ready. The documentation is ahead of the released code here.

You cannot failover in a WAN setup with VPN. The latencies are too high to setup a HA cluster.

Du hast auch nur zwei OSDs in der Device-Class "large_ssd". Wie soll das funktionieren?

Because that's a single client. As you see with multiple clients performance scales up well.

Sorry, sometimes the Ceph documentation describes features that have not been released yet.

I would replace "firstn 3" with "firstn 0" to make the rules more generic. I do not if it's possible to remove the global stretch mode from a cluster. Maybe you should ask on the Ceph mailing list.

You have enabled stretch mode for the whole cluster. This means that all PGs from all pools need to be distributed across both DCs before they become active. Quoting the docs https://docs.ceph.com/en/squid/rados/operations/stretch-mode/ : What your want can be achieved with enabling stretch...