2

I have a situation where my OSDs are full

1 full osd(s) 2 nearfull osd(s) 2 pool(s) full

and deleting the PVC is just stuck in "terminating". I tried the trick of forcing finalizers to null, but that still leaves data around. Next time I create a PVC I cannot create any new data because OSDs and pools are full.

How do I delete all data that is on OSD without compromising the ceph cluster. All I have there is test data and nothing of value.

user3304297
  • 131
  • 2
  • 8
  • Which version of Kubernetes did you use? – kkopczak Nov 12 '21 at 14:48
  • 1
    I found [this manual](https://github.com/rook/rook/blob/master/Documentation/ceph-osd-mgmt.md#purge-the-osd-manually), does that help? There is also some kind of automation just [above](https://github.com/rook/rook/blob/master/Documentation/ceph-osd-mgmt.md#purge-the-osd-from-the-ceph-cluster). If not - could you please add more exact details/errors? – kkopczak Nov 15 '21 at 19:03

0 Answers0