Add support for availability zones to the crush map

If the deployemnt environment supports it, and the user would like to,
we can use Juju's Availability Zone information to setup the Ceph
cluster to use Availability Zones as the failure domain instead of
the host.

Change-Id: I4566696750b388918761ded0ed5beb0bf82ff501
Depends-On: Ie25ac1b001db558d6a40fe3eaca014e8f4174241

config.yaml

@@ -104,6 +104,12 @@ options:
       threads-max to a high value to avoid problems with large numbers (>20)
       of OSDs recovering. very large clusters should set those values even
       higher (e.g. max for kernel.pid_max is 4194303).
+  customize-failure-domain:
+    type: boolean
+    default: false
+    description: |
+      Setting this to true will tell Ceph to replicate across Juju's
+      Availability Zone instead of specifically by host.
   nagios_context:
     type: string
     default: "juju"

hooks/ceph_hooks.py

@@ -417,6 +417,33 @@ def mon_relation():
         ceph.bootstrap_monitor_cluster(leader_get('monitor-secret'))
         ceph.wait_for_bootstrap()
         ceph.wait_for_quorum()
+        # If we can and want to
+        if is_leader() and config('customize-failure-domain'):
+            # But only if the environment supports it
+            if os.environ.get('JUJU_AVAILABILITY_ZONE'):
+                cmds = [
+                    "ceph osd getcrushmap -o /tmp/crush.map",
+                    "crushtool -d /tmp/crush.map| "
+                    "sed 's/step chooseleaf firstn 0 type host/step "
+                    "chooseleaf firstn 0 type rack/' > "
+                    "/tmp/crush.decompiled",
+                    "crushtool -c /tmp/crush.decompiled -o /tmp/crush.map",
+                    "crushtool -i /tmp/crush.map --test",
+                    "ceph osd setcrushmap -i /tmp/crush.map"
+                ]
+                for cmd in cmds:
+                    try:
+                        subprocess.check_call(cmd, shell=True)
+                    except subprocess.CalledProcessError as e:
+                        log("Failed to modify crush map:", level='error')
+                        log("Cmd: {}".format(cmd), level='error')
+                        log("Error: {}".format(e.output), level='error')
+                        break
+            else:
+                log(
+                    "Your Juju environment doesn't"
+                    "have support for Availability Zones"
+                )
         notify_osds()
         notify_radosgws()
         notify_client()