This PR removes the command to change the replication size to 1.
Now that we can have OSDs backed by loop files, this setting
doesn't make sense anymore.
Signed-off-by: Luciano Lo Giudice <email address hidden>
6556b0e...
by
hemanthnakkina <email address hidden>
Merge pull request #11 from gnuoy/remove-default-crush-rule
A recent change to microceph *1 caused the default crush rule called
`replicated_rule` to be removed. Instead a rule called
`microceph_auto_osd` is created on single node deployments. If/when
the cluster is expanded sufficently a new rule called
`microceph_auto_host` is created automatically and existing pools
migrated to use this, all this is done automatically be microceph.
This change stop a profile being set at all to take advantage of this
new behaviour.
Currently already treated broker requests check
in store for request ids. However the data source
should be the unit relation data. Change the logic
to check if the request is already processed in
unit relation data.
Add mon_allow_pool_size_one and osd_pool_default_size
to ceph configuration so that ceph supports single
node deployment.
Use --yes-i-really-mean-it flag while setting the pool
size to 1.