Ceph RBD mirroring

Overview

RADOS Block Device (RBD) mirroring is a process of asynchronous replication of Ceph block device images between two or more Ceph clusters. Mirroring ensures point-in-time consistent replicas of all changes to an image, including reads and writes, block device resizing, snapshots, clones, and flattening. RBD mirroring is mainly used for disaster recovery (i.e. having a secondary site as a failover). See Upstream Ceph documentation on RBD mirroring for complete information.

This guide will show how to deploy two Ceph clusters with RBD mirroring between them with the use of the ceph-rbd-mirror charm. See the charm’s documentation for basic information and charm limitations.

RBD mirroring is only one aspect of datacentre redundancy. Refer to Ceph RADOS Gateway Multisite Replication and other work to arrive at a complete solution.

Note

RBD mirroring makes use of the journaling feature of Ceph. This incurs an overhead for write activity on an RBD image that will adversely affect performance. See Florian Haas’ performance analysis of RBD mirror from Cephalocon Barcelona 2019.

Requirements

The two Ceph clusters will correspond to sites ‘a’ and ‘b’ and each cluster will reside within a separate model (models ‘site-a’ and ‘site-b’). The deployment will require the use of Cross model relations.

Deployment characteristics:

  • each cluster will have 7 units:

    • 3 x ceph-osd

    • 3 x ceph-mon

    • 1 x ceph-rbd-mirror

  • application names will be used to distinguish between applications in site ‘a’ from those in site ‘b’ (e.g. ceph-mon-a and ceph-mon-b)

  • the ceph-osd units will use block device /dev/vdd for their OSD volumes

Note

The two Ceph clusters can optionally be placed within the same model, and thus obviate the need for cross model relations. This topology is not generally considered to be a real world scenario.

Deployment

For site ‘a’ the following configuration is placed into file site-a.yaml:

ceph-mon-a:
  monitor-count: 3
  expected-osd-count: 3
  source: distro

ceph-osd-a:
  osd-devices: /dev/vdd
  source: distro

ceph-rbd-mirror-a:
  source: distro

Create the model and deploy the software for each site:

  • Site ‘a’

    juju add-model site-a
    juju deploy -n 3 --config site-a.yaml ceph-osd ceph-osd-a
    juju deploy -n 3 --config site-a.yaml ceph-mon ceph-mon-a
    juju deploy --config site-a.yaml ceph-rbd-mirror ceph-rbd-mirror-a
    
  • Site ‘b’

    An analogous configuration file is used (i.e. replace ‘a’ with ‘b’):

    juju add-model site-b
    juju deploy -n 3 --config site-b.yaml ceph-osd ceph-osd-b
    juju deploy -n 3 --config site-b.yaml ceph-mon ceph-mon-b
    juju deploy --config site-b.yaml ceph-rbd-mirror ceph-rbd-mirror-b
    

Add two local relations for each site:

  • Site ‘a’

    juju add-relation -m site-a ceph-mon-a:osd ceph-osd-a:mon
    juju add-relation -m site-a ceph-mon-a:rbd-mirror ceph-rbd-mirror-a:ceph-local
    
  • Site ‘b’

    juju add-relation -m site-b ceph-mon-b:osd ceph-osd-b:mon
    juju add-relation -m site-b ceph-mon-b:rbd-mirror ceph-rbd-mirror-b:ceph-local
    

Export a ceph-rbd-mirror endpoint (by means of an “offer”) for each site. This will enable us to create the inter-site (cross model) relations:

  • Site ‘a’

    juju switch site-a
    juju offer ceph-rbd-mirror-a:ceph-remote
    

    Output:

    Application "ceph-rbd-mirror-a" endpoints [ceph-remote] available at "admin/site-a.ceph-rbd-mirror-a"
    
  • Site ‘b’

    juju switch site-b
    juju offer ceph-rbd-mirror-b:ceph-remote
    

    Output:

    Application "ceph-rbd-mirror-b" endpoints [ceph-remote] available at "admin/site-b.ceph-rbd-mirror-b"
    

Add the two inter-site relations by referring to the offer URLs (included in the output above) as if they were applications in the local model:

juju add-relation -m site-a ceph-mon-a admin/site-b.ceph-rbd-mirror-b
juju add-relation -m site-b ceph-mon-b admin/site-a.ceph-rbd-mirror-a

Verify the output of juju status for each model:

juju status -m site-a

Output:

Model   Controller   Cloud/Region    Version  SLA          Timestamp
site-a  maas-prod-1  acme-1/default  2.8.1    unsupported  20:00:41Z

SAAS               Status   Store        URL
ceph-rbd-mirror-b  waiting  icarus-maas  admin/site-b.ceph-rbd-mirror-b

App                Version  Status   Scale  Charm            Store       Rev  OS      Notes
ceph-mon-a         15.2.3   active       3  ceph-mon         jujucharms   49  ubuntu
ceph-osd-a         15.2.3   active       3  ceph-osd         jujucharms  304  ubuntu
ceph-rbd-mirror-a  15.2.3   waiting      1  ceph-rbd-mirror  jujucharms   12  ubuntu

Unit                  Workload  Agent  Machine  Public address  Ports  Message
ceph-mon-a/0*         active    idle   0/lxd/0  10.0.0.57              Unit is ready and clustered
ceph-mon-a/1          active    idle   1/lxd/0  10.0.0.58              Unit is ready and clustered
ceph-mon-a/2          active    idle   2/lxd/0  10.0.0.59              Unit is ready and clustered
ceph-osd-a/0*         active    idle   0        10.0.0.69              Unit is ready (1 OSD)
ceph-osd-a/1          active    idle   1        10.0.0.19              Unit is ready (1 OSD)
ceph-osd-a/2          active    idle   2        10.0.0.20              Unit is ready (1 OSD)
ceph-rbd-mirror-a/0*  waiting   idle   3        10.0.0.22              Waiting for pools to be created

Machine  State    DNS        Inst id              Series  AZ       Message
0        started  10.0.0.69  virt-node-08         focal   default  Deployed
0/lxd/0  started  10.0.0.57  juju-bb0dc1-0-lxd-0  focal   default  Container started
1        started  10.0.0.19  virt-node-10         focal   default  Deployed
1/lxd/0  started  10.0.0.58  juju-bb0dc1-1-lxd-0  focal   default  Container started
2        started  10.0.0.20  virt-node-11         focal   default  Deployed
2/lxd/0  started  10.0.0.59  juju-bb0dc1-2-lxd-0  focal   default  Container started
3        started  10.0.0.22  virt-node-03         focal   default  Deployed

Offer              Application        Charm            Rev  Connected  Endpoint     Interface        Role
ceph-rbd-mirror-a  ceph-rbd-mirror-a  ceph-rbd-mirror  12   1/1        ceph-remote  ceph-rbd-mirror  requirer
juju status -m site-b

Output:

Model   Controller   Cloud/Region    Version  SLA          Timestamp
site-b  maas-prod-1  acme-1/default  2.8.1    unsupported  20:02:58Z

SAAS               Status   Store        URL
ceph-rbd-mirror-a  waiting  icarus-maas  admin/site-a.ceph-rbd-mirror-a

App                Version  Status   Scale  Charm            Store       Rev  OS      Notes
ceph-mon-b         15.2.3   active       3  ceph-mon         jujucharms   49  ubuntu
ceph-osd-b         15.2.3   active       3  ceph-osd         jujucharms  304  ubuntu
ceph-rbd-mirror-b  15.2.3   waiting      1  ceph-rbd-mirror  jujucharms   12  ubuntu

Unit                  Workload  Agent  Machine  Public address  Ports  Message
ceph-mon-b/0*         active    idle   0/lxd/0  10.0.0.60              Unit is ready and clustered
ceph-mon-b/1          active    idle   1/lxd/0  10.0.0.61              Unit is ready and clustered
ceph-mon-b/2          active    idle   2/lxd/0  10.0.0.62              Unit is ready and clustered
ceph-osd-b/0*         active    idle   0        10.0.0.21              Unit is ready (1 OSD)
ceph-osd-b/1          active    idle   1        10.0.0.54              Unit is ready (1 OSD)
ceph-osd-b/2          active    idle   2        10.0.0.55              Unit is ready (1 OSD)
ceph-rbd-mirror-b/0*  waiting   idle   3        10.0.0.56              Waiting for pools to be created

Machine  State    DNS        Inst id              Series  AZ       Message
0        started  10.0.0.21  virt-node-02         focal   default  Deployed
0/lxd/0  started  10.0.0.60  juju-3ef7c5-0-lxd-0  focal   default  Container started
1        started  10.0.0.54  virt-node-04         focal   default  Deployed
1/lxd/0  started  10.0.0.61  juju-3ef7c5-1-lxd-0  focal   default  Container started
2        started  10.0.0.55  virt-node-05         focal   default  Deployed
2/lxd/0  started  10.0.0.62  juju-3ef7c5-2-lxd-0  focal   default  Container started
3        started  10.0.0.56  virt-node-06         focal   default  Deployed

Offer              Application        Charm            Rev  Connected  Endpoint     Interface        Role
ceph-rbd-mirror-b  ceph-rbd-mirror-b  ceph-rbd-mirror  12   1/1        ceph-remote  ceph-rbd-mirror  requirer

There are no Ceph pools created by default. The next section (‘Pool creation’) provides guidance.

Pool creation

RBD pools can be created by either a supporting charm (through the Ceph broker protocol) or manually by the operator:

  1. A charm-created pool (e.g. the glance or nova-compute charms) will automatically be detected and acted upon (i.e. a remote pool will be set up in the peer cluster).

  2. A manually-created pool, whether done via the ceph-mon application or through Ceph directly, will require an action to be run on the ceph-rbd-mirror application leader in order for the remote pool to come online.

    For example, to create a pool manually in site ‘a’ and have ceph-rbd-mirror (of site ‘a’) initialise a pool in site ‘b’:

    juju run-action --wait -m site-a ceph-mon-a/leader create-pool name=mypool app-name=rbd
    juju run-action --wait -m site-a ceph-rbd-mirror-a/leader refresh-pools
    

    This can be verified by listing the pools in site ‘b’:

    juju run-action --wait -m site-b ceph-mon-b/leader list-pools
    

Note

Automatic peer-pool creation (for a charm-created pool) is based on the local pool being labelled with a Ceph ‘rbd’ tag. This Ceph-internal labelling occurs when the newly-created local pool is associated with the RBD application. This last feature is supported starting with Ceph Luminous (OpenStack Queens).

Failover and fallback

To manage failover and fallback, the demote and promote actions are applied to the ceph-rbd-mirror application leader.

For instance, to fail over from site ‘a’ to site ‘b’ the former is demoted and the latter is promoted. The rest of the commands are status checks:

juju run-action --wait -m site-a ceph-rbd-mirror-a/leader status verbose=true
juju run-action --wait -m site-b ceph-rbd-mirror-b/leader status verbose=true

juju run-action --wait -m site-a ceph-rbd-mirror-a/leader demote

juju run-action --wait -m site-a ceph-rbd-mirror-a/leader status verbose=true
juju run-action --wait -m site-b ceph-rbd-mirror-b/leader status verbose=true

juju run-action --wait -m site-b ceph-rbd-mirror-b/leader promote

To fall back to site ‘a’ the actions are reversed:

juju run-action --wait -m site-b ceph-rbd-mirror-b/leader demote
juju run-action --wait -m site-a ceph-rbd-mirror-a/leader promote

Note

With Ceph Luminous (and greater), the mirror status information may not be accurate. Specifically, the entries_behind_master counter may never get to ‘0’ even though the image has been fully synchronised.

Recovering from abrupt shutdown

It is possible that an abrupt shutdown and/or an interruption to communication channels may lead to a “split-brain” condition. This may cause the mirroring daemon in each cluster to claim to be the primary. In such cases, the operator must make a call as to which daemon is correct. Generally speaking, this means deciding which cluster has the most recent data.

Elect a primary by applying the demote and promote actions to the appropriate ceph-rbd-mirror leader. After doing so, the resync-pools action must be run on the secondary cluster leader. The promote action may require a force option.

Here, we make site ‘a’ be the primary by demoting site ‘b’ and promoting site ‘a’:

juju run-action --wait -m site-b ceph-rbd-mirror/leader demote
juju run-action --wait -m site-a ceph-rbd-mirror/leader promote force=true

juju run-action --wait -m site-a ceph-rbd-mirror/leader status verbose=true
juju run-action --wait -m site-b ceph-rbd-mirror/leader status verbose=true

juju run-action --wait -m site-b ceph-rbd-mirror/leader resync-pools i-really-mean-it=true

Note

When using Ceph Luminous, the mirror state information will not be accurate after recovering from unclean shutdown. Regardless of the output of the status information, you will be able to write to images after a forced promote.