1

I have a 3 node ceph proxmox cluster for storage with node1 7*143GB ssds, node2 7*143 ssds, node3 7*218 ssds. Now I have read through the http://docs.ceph.com/docs/jewel/architecture/#mapping-pgs-to-osds however it is not 100% clear to me if I can create a VM with hard disk size 300GB using Ceph pool. Please note that largest single ssd i have is 218GB big. From the documentation, it looks like it stores objects in placement groups and placement groups in separate nodes and drives. If that is the case then I want to say, it shouldn't be a problem.

Does anyone has insight on this? Thanks in advance.

fcukinyahoo
  • 145
  • 1
  • 1
  • 6

1 Answers1

1

Concerning your question, it is true that it won't be a problem unless that VM is actually trying to take the more than 218GB from the same ssd.

Yes, all the placements groups are stored in separate nodes and in your case you can create a VM , but make sure the HD size should be dynamic so that dynamically it can eat up the space when required.

I would recommend you to put some alerts / monitoring on the SSD aat around 85% so that , when it will hit 85% you can quickly revise that and take immediate action on this.

Ranadip Dutta
  • 123
  • 8
  • 2
    Thanks for the reply. Yes all VMs hard disk sizes grow dynamically. After creating the VM, i checked Ceph and it did distribute according to pg_num so all seems good. Thanks. – fcukinyahoo Nov 30 '16 at 15:58