ceph-iscsi does not provision additional gateway when a new unit of ceph-iscsi is deployed
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Ceph iSCSI Charm |
New
|
Undecided
|
Unassigned |
Bug Description
Consider the following scenario.
Two ceph-iscsi units are deployed. The machine hosting one unit irreversibly fails and has to be replaced. The unit is lost.
Now we want to provision additional ceph-iscsi unit to fill in for the lost one.
From the user perspective I'd expect that the charm automatically handles adding a new node as a gateway.
Currently it does not happen. The operator must run manual actions with `gwcli` utility to remove failed gateway and add the new one:
sudo gwcli
/> cd iscsi-targets/
/iscsi-
o- gateways .......
o- node02.maas .......
o- node03.maas .......
/iscsi-
Deleting gateway, node03.maas
/iscsi-
OS version/package checks have been bypassed
Adding gateway, sync'ing 1 disk(s) and 1 client(s)
All gateways accessible
ok