a2455bbe5a
* Fixed typo in test_ceph_iscsi_charm.py file name * Replaced artifical emit in test_on_has_peers with by adding relation data later in the test. * Updated calls to `add_relation_unit` as the `remote_unit_data` arg has been removed. |
||
---|---|---|
hooks | ||
mod | ||
src | ||
templates | ||
tests | ||
unit_tests | ||
.gitignore | ||
.gitmodules | ||
.stestr.conf | ||
actions.yaml | ||
charm-prep.sh | ||
config.yaml | ||
metadata.yaml | ||
README.md | ||
requirements.txt | ||
test-requirements.txt | ||
todo.txt | ||
tox.ini |
Overview
The charm provides the Ceph iSCSI gateway service. It is intended to be used in conjunction with the ceph-osd and ceph-mon charms.
Warning
: This charm is in a preview state for testing and should not be used outside of the lab.
Usage
Deployment
When deploying ceph-iscsi ensure that exactly two units of the charm are being deployed, this will provide multiple data paths to clients.
Note
: Deploying four units is also theoretical possible but has not been tested.
A sample bundle.yaml
file's contents:
series: focal
machines:
'0':
'1':
'2':
applications:
ceph-iscsi:
charm: cs:ceph-iscsi
num_units: 2
to:
- lxd:0
- lxd:1
ceph-osd:
charm: cs:ceph-osd
num_units: 3
storage:
osd-devices: /dev/vdb
to:
- '0'
- '1'
- '2'
ceph-mon:
charm: cs:ceph-mon
num_units: 3
options:
monitor-count: '3'
to:
- lxd:0
- lxd:1
- lxd:2
relations:
- - ceph-mon:client
- ceph-iscsi:ceph-client
- - ceph-osd:mon
- ceph-mon:osd
Important
: Make sure the designated block device passed to the ceph-osd charms exists and is not currently in use.
Deploy the bundle:
juju deploy ./bundle.yaml
Managing Targets
The charm provides an action for creating a simple target. If more complex
managment of targets is requires then the gwcli
tool should be used. gwcli
is available from the root account on the gateway nodes.
$ juju ssh ceph-iscsi/1
$ sudo gwcli
/> ls
Actions
This section covers Juju actions supported by the charm. Actions allow specific operations to be performed on a per-unit basis.
create-target
Run this action to create an iscsi target.
$ juju run-action ceph-iscsi/0 create-target \
image-size=2G \
image-name=bob \
pool-name=superssd \
client-initiatorname=iqn.1993-08.org.debian:01:aaa2299be916 \
client-username=usera \
client-password=testpass
Action queued with id: "28"
If the iqn of the created target is returned in the ouput from the action:
$ juju show-action-output 28
UnitId: ceph-iscsi/0
results:
iqn: iqn.2003-01.com.ubuntu.iscsi-gw:iscsi-igw
status: completed
timing:
completed: 2020-04-02 13:32:02 +0000 UTC
enqueued: 2020-04-02 13:18:42 +0000 UTC
started: 2020-04-02 13:18:45 +0000 UTC
pause
Pause the ceph-iscsi unit. This action will stop the rbd services.
resume
Resume the ceph-iscsi unit. This action will start the rbd services if paused.
Network spaces
This charm supports the use of Juju network spaces (Juju
v.2.0
). This feature optionally allows specific types of the application's
network traffic to be bound to subnets that the underlying hardware is
connected to.
Note
: Spaces must be configured in the backing cloud prior to deployment.
The ceph-iscsi charm exposes the following traffic types (bindings):
- 'public' (front-side)
- 'cluster' (back-side)
For example, providing that spaces 'data-space' and 'cluster-space' exist, the deploy command above could look like this:
juju deploy --config ceph-iscsi.yaml -n 2 ceph-iscsi \
--bind "public=data-space cluster=cluster-space"
Alternatively, configuration can be provided as part of a bundle:
ceph-iscsi:
charm: cs:ceph-iscsi
num_units: 2
bindings:
public: data-space
cluster: cluster-space
VMWare integration
-
Create ceph pool if required.
To create a new pool to back the iscsi targets run the create-pool action from the ceph-mon charm.
$ juju run-action ceph-mon/0 create-pool name=iscsi-targets
Action queued with id: "1"
$ juju show-action-output 1
UnitId: ceph-mon/0
results:
Stderr: |
pool 'iscsi-targets' created
set pool 2 size to 3
set pool 2 target_size_ratio to 0.1
enabled application 'unknown' on pool 'iscsi-targets'
set pool 2 pg_autoscale_mode to on
status: completed
timing:
completed: 2020-04-08 06:42:00 +0000 UTC
enqueued: 2020-04-08 06:41:38 +0000 UTC
started: 2020-04-08 06:41:42 +0000 UTC
-
Collect the Initiator name for adapter.
From the VMWare admin UI select the
Adapters
tab in the Storage context. EnsureiSCSI enabled
is set toEnabled
.Click 'Configure iSCSI' and take a note of the
iqn
name. -
Create iSCSI target.
Run the action to create a target for VMWare to use.
Note
: The username should be more than eight characters and the password between twelve and sixteen characters.
$ juju run-action ceph-iscsi/0 create-target \
client-initiatorname="iqn.1998-01.com.vmware:node-caloric-02f98bac" \
client-username=vmwareclient \
client-password=12to16characters \
image-size=10G \
image-name=disk_1 \
pool-name=iscsi-targets
$ juju show-action-output 2
UnitId: ceph-iscsi/0
results:
Stdout: |
Warning: Could not load preferences file /root/.gwcli/prefs.bin.
iqn: iqn.2003-01.com.ubuntu.iscsi-gw:iscsi-igw
status: completed
timing:
completed: 2020-04-08 06:58:34 +0000 UTC
enqueued: 2020-04-08 06:58:15 +0000 UTC
started: 2020-04-08 06:58:19 +0000 UTC
-
Add target to VMWare.
Follow the Ceph iSCSI Gateway documentation to use the new target. Use CHAP username and password provided to the
create-target
action.
Warning
: As of the time of writing the workaround to set the CHAP credentials via the esx cli is still needed.