From 24e62faa2d4652ac2fdb6c6f12dc3e3379697463 Mon Sep 17 00:00:00 2001 From: Satoru Takeuchi Date: Thu, 23 Jan 2020 01:07:35 +0000 Subject: [PATCH] ceph: add a note of osd in placement Signed-off-by: Satoru Takeuchi --- Documentation/ceph-cluster-crd.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/Documentation/ceph-cluster-crd.md b/Documentation/ceph-cluster-crd.md index 0022acb23..50c91ef9c 100755 --- a/Documentation/ceph-cluster-crd.md +++ b/Documentation/ceph-cluster-crd.md @@ -311,6 +311,8 @@ A Placement configuration is specified (according to the kubernetes PodSpec) as: The `mon` pod does not allow `Pod` affinity or anti-affinity. Instead, `mon`s have built-in anti-affinity with each other through the operator. The operator determines which nodes should run a `mon`. Each `mon` is then tied to a node with a node selector using a hostname. See the [mon design doc](https://github.com/rook/rook/blob/master/design/ceph/mon-health.md) for more details on the `mon` failover design. +If you use `labelSelector` for `osd` pods, you must write two rules both for `rook-ceph-osd` and `rook-ceph-osd-prepare` like [the example configuration](https://github.com/rook/rook/blob/master/cluster/examples/kubernetes/ceph/cluster-on-pvc.yaml#L68). It comes from the design that there are these two pods for an OSD. For more detail, see the [osd design doc](https://github.com/rook/rook/blob/master/design/ceph/dedicated-osd-pod.md) and [the related issue](https://github.com/rook/rook/issues/4582). + The Rook Ceph operator creates a Job called `rook-ceph-detect-version` to detect the full Ceph version used by the given `cephVersion.image`. The placement from the `mon` section is used for the Job. ### Cluster-wide Resources Configuration Settings