Merge lp:~gandelman-a/charms/precise/hacluster/ocf into lp:charms/hacluster

[Adam Gandelman]

This ships the ceph RBD resource agent directly in the charm, copied from the version of ceph shipped /w quantal (.48) This is a complicated matter currently and hopefully this is only a temporary solution:

- The .46 ceph-resource-agents in precise do not include the rbd RA.
- The .48 ceph-resource-agents in quantal does. However, setting up cloud-archive access to install it would also upgrade any ceph packages installed by the principle.
- The .56 ceph-resource-agents in raring is not compatible /w the rbd CLI in .46 or .48.

I'd like to ping the OCF author about possibly supporting different versions /w a single RA.

Until then, ACAICS best solution is to ship and install RA's /w the charm.

[James Page (james-page)]

Adam - generally looks OK;

1) Ceph resource agents still gets installed - think we need to drop this right?

    if True in [ra.startswith('ocf:ceph')
                for ra in resources.itervalues()]:
        utils.install('ceph-resource-agents')

Other than that look good to me.

review: Needs Fixing

[Adam Gandelman (gandelman-a)]

The rbd OCF is not included in the .041 release, but it is /w .48 + .56. The .56 OCF is updated for rbd syntax changes. I thought the easiest way to keep this charm able to deploy across all versions would be to install the rbd OCF packaged with the charm along with the package, so that if it is included /w the package proper, it will override the one copied from the charm.

[James Page (james-page)]

OK - testing with the mysql and rabbitmq updates which use these features - will verify on raring and precise.

=== modified file 'hooks/hooks.py'
--- hooks/hooks.py	2013-02-20 10:10:42 +0000
+++ hooks/hooks.py	2013-03-02 03:56:20 +0000
@@ -7,6 +7,7 @@
 #  Andres Rodriguez <[e-mail]>
 #
 
+import shutil
 import sys
 import time
 import os
@@ -20,10 +21,17 @@
     utils.juju_log('INFO', 'Begin install hook.')
     utils.configure_source()
     utils.install('corosync', 'pacemaker', 'python-netaddr', 'ipmitool')
+    # XXX rbd OCF only included with newer versions of ceph-resource-agents.
+    # Bundle /w charm until we figure out a better way to install it.
+    if not os.path.exists('/usr/lib/ocf/resource.d/ceph'):
+        os.makedirs('/usr/lib/ocf/resource.d/ceph')
+    if not os.path.isfile('/usr/lib/ocf/resource.d/ceph/rbd'):
+        shutil.copy('ocf/ceph/rbd', '/usr/lib/ocf/resource.d/ceph/rbd')
     utils.juju_log('INFO', 'End install hook.')
 
 
 def get_corosync_conf():
+    conf = {}
     for relid in utils.relation_ids('ha'):
         for unit in utils.relation_list(relid):
             conf = {
@@ -39,6 +47,9 @@
                 }
             if None not in conf.itervalues():
                 return conf
+    missing = [k for k, v in conf.iteritems() if v == None]
+    utils.juju_log('INFO',
+                   'Missing required principle configuration: %s' % missing)
     return None
 
 

=== added directory 'ocf'
=== added directory 'ocf/ceph'
=== added file 'ocf/ceph/rbd'
--- ocf/ceph/rbd	1970-01-01 00:00:00 +0000
+++ ocf/ceph/rbd	2013-03-02 03:56:20 +0000
@@ -0,0 +1,296 @@
+#!/bin/sh
+#
+#   OCF resource agent for mapping and unmapping
+#   RADOS Block Devices (RBDs)
+#
+#   License:      GNU Lesser General Public License (LGPL) 2.1
+#   (c) 2012 Florian Haas, hastexo
+#
+
+# Initialization:
+: ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat}
+. ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs
+
+# Convenience variables
+# When sysconfdir isn't passed in as a configure flag,
+# it's defined in terms of prefix
+prefix=/usr
+
+# Defaults
+OCF_RESKEY_pool_default="rbd"
+OCF_RESKEY_cephconf_default="/etc/ceph/ceph.conf"
+: ${OCF_RESKEY_pool=${OCF_RESKEY_pool_default}}
+: ${OCF_RESKEY_cephconf=${OCF_RESKEY_cephconf_default}}
+
+rbd_meta_data() {
+    cat <<EOF
+<?xml version="1.0"?>
+<!DOCTYPE resource-agent SYSTEM "ra-api-1.dtd">
+<resource-agent name="rbd" version="0.1">
+  <version>0.1</version>
+  <longdesc lang="en">
+Manages RADOS Block Devices (RBDs) as a highly available
+resource. Maps and unmaps RBDs as needed.
+  </longdesc>
+  <shortdesc lang="en">Maps and unmaps RADOS Block Devices</shortdesc>
+  <parameters>
+    <parameter name="name" unique="0" required="1">
+      <longdesc lang="en">
+      Name of the RBD device.
+      </longdesc>
+      <shortdesc lang="en">RBD device name</shortdesc>
+      <content type="string"/>
+    </parameter>
+    <parameter name="pool" unique="0" required="0">
+      <longdesc lang="en">
+      Name of the RADOS pool where the RBD has been created
+      </longdesc>
+      <shortdesc lang="en">RADOS pool name</shortdesc>
+      <content type="string" default="${OCF_RESKEY_pool_default}"/>
+    </parameter>
+    <parameter name="snap" unique="0" required="0">
+      <longdesc lang="en">
+      Name of the device snapshot to map.
+      </longdesc>
+      <shortdesc lang="en">Snapshot name</shortdesc>
+      <content type="string"/>
+    </parameter>
+    <parameter name="cephconf" unique="0" required="0">
+      <longdesc lang="en">
+      Location of the Ceph configuration file
+      </longdesc>
+      <shortdesc lang="en">Ceph configuration file</shortdesc>
+      <content type="string" default="${OCF_RESKEY_cephconf_default}"/>
+    </parameter>
+    <parameter name="mon" unique="0" required="0">
+      <longdesc lang="en">
+      Address (or comma-separated list of addresses) of
+      monitor servers to connect to. Overrides values from
+      configuration file.
+      </longdesc>
+      <shortdesc lang="en">Monitor address(es)</shortdesc>
+      <content type="string"/>
+    </parameter>
+    <parameter name="user" unique="0" required="0">
+      <longdesc lang="en">
+      Username to use when mapping the device. Required
+      if Ceph authentication is enabled on the monitor.
+      </longdesc>
+      <shortdesc lang="en">Authentication username</shortdesc>
+      <content type="string"/>
+    </parameter>
+    <parameter name="secret" unique="0" required="0">
+      <longdesc lang="en">
+      File containing an authentication secret. Required
+      if Ceph authentication is enabled on the monitor.
+      </longdesc>
+      <shortdesc lang="en">Authentication secret file</shortdesc>
+      <content type="string"/>
+    </parameter>
+  </parameters>
+  <actions>
+    <action name="start"        timeout="20" />
+    <action name="stop"         timeout="20" />
+    <action name="monitor"      timeout="20"
+                                interval="10" depth="0" />
+    <action name="meta-data"    timeout="5" />
+    <action name="validate-all"   timeout="20" />
+  </actions>
+</resource-agent>
+EOF
+}
+
+rbd_usage() {
+        cat <<EOF
+usage: $0 {start|stop|status|monitor|validate-all|meta-data}
+
+Expects to have a fully populated OCF RA-compliant environment set.
+EOF
+}
+
+# rbd command wrapper: builds an option string for invoking RBD based
+# on resource parameters, and invokes rbd through ocf_run.
+do_rbd() {
+    local rbd_options
+
+    if [ -n "${OCF_RESKEY_cephconf}" ]; then
+        rbd_options="$rbd_options -c ${OCF_RESKEY_cephconf}"
+    fi
+    if [ -n "${OCF_RESKEY_mon}" ]; then
+        rbd_options="$rbd_options -m ${OCF_RESKEY_mon}"
+    fi
+
+    ocf_run rbd $rbd_options $@
+}
+
+# Convenience function that uses "rbd showmapped" to retrieve the
+# mapped device name from the pool, RBD name, and snapshot.
+find_rbd_dev() {
+    local sedpat
+
+    # Example output from "rbd showmapped" (tab separated):
+    # id        pool    image   snap    device
+    # 0         rbd     test    -       /dev/rbd0
+
+    # Build the sed pattern, substituting "-" for the snapshot name if
+    # it's unset
+    sedpat="[0-9]\+\t${OCF_RESKEY_pool}\t${OCF_RESKEY_name}\t${OCF_RESKEY_snap:--}\t\(/dev/rbd[0-9]\+\)"
+
+    # Run rbd showmapped, filter out the header line, then try to
+    # extract the device name
+    rbd showmapped | tail -n +2 | sed -n -e "s,$sedpat,\1,p"
+}
+
+rbd_validate_all() {
+    # Test for configuration errors first
+    if [ -z "$OCF_RESKEY_name" ]; then
+       ocf_log err 'Required parameter "name" is unset!'
+       exit $OCF_ERR_CONFIGURED
+    fi
+
+    # Test for required binaries
+    check_binary rbd
+
+    return $OCF_SUCCESS
+}
+
+rbd_monitor() {
+    local rc
+    local rbd_dev
+
+    if ! [ -d /sys/bus/rbd ]; then
+        ocf_log debug "rbd module is not loaded"
+        return $OCF_NOT_RUNNING
+    fi
+
+    rbd_dev=`find_rbd_dev`
+
+    if [ -z "$rbd_dev" ]; then
+        ocf_log debug "RBD device is unmapped"
+        rc=$OCF_NOT_RUNNING
+    elif [ -b "$rbd_dev" ]; then
+        ocf_log debug "RBD device is mapped to $rbd_dev"
+        rc=$OCF_SUCCESS
+    else
+        # Device is listed, but the corresponding path is not a block
+        # device.
+        ocf_log err "$rbd_dev is not a block device!"
+        rc=$OCF_ERR_GENERIC
+    fi
+
+    return $rc
+}
+
+rbd_start() {
+    local rbd_map_options
+    local rbd_name
+
+    # if resource is already running, bail out early
+    if rbd_monitor; then
+        ocf_log info "Resource is already running"
+        return $OCF_SUCCESS
+    fi
+
+    # actually start up the resource here (make sure to immediately
+    # exit with an $OCF_ERR_ error code if anything goes seriously
+    # wrong)
+
+    if [ ! -d /sys/bus/rbd ]; then
+        ocf_run modprobe -v rbd || exit $OCF_ERR_INSTALLED
+    fi
+
+    if [ -n "${OCF_RESKEY_user}" ]; then
+        rbd_map_options="--user ${OCF_RESKEY_user}"
+    fi
+    if [ -n "${OCF_RESKEY_secret}" ]; then
+        rbd_map_options="$rbd_map_options --secret ${OCF_RESKEY_secret}"
+    fi
+
+    rbd_name="${OCF_RESKEY_pool}/${OCF_RESKEY_name}"
+    if [ -n "${OCF_RESKEY_snap}" ]; then
+        rbd_name="$rbd_name@${OCF_RESKEY_snap}"
+    fi
+
+    do_rbd map $rbd_name $rbd_map_options || exit $OCF_ERR_GENERIC
+
+    # After the resource has been started, check whether it started up
+    # correctly. If the resource starts asynchronously, the agent may
+    # spin on the monitor function here -- if the resource does not
+    # start up within the defined timeout, the cluster manager will
+    # consider the start action failed
+    while ! rbd_monitor; do
+        ocf_log debug "Resource has not started yet, waiting"
+        sleep 1
+    done
+
+    # only return $OCF_SUCCESS if _everything_ succeeded as expected
+    return $OCF_SUCCESS
+}
+
+rbd_stop() {
+    local rc
+    local rbd_dev
+
+    rbd_monitor
+    rc=$?
+    case "$rc" in
+        "$OCF_SUCCESS")
+            # Currently running. Normal, expected behavior.
+            ocf_log debug "Resource is currently running"
+            ;;
+        "$OCF_NOT_RUNNING")
+            # Currently not running. Nothing to do.
+            ocf_log info "Resource is already stopped"
+            return $OCF_SUCCESS
+            ;;
+    esac
+
+    # actually shut down the resource here (make sure to immediately
+    # exit with an $OCF_ERR_ error code if anything goes seriously
+    # wrong)
+    rbd_dev=`find_rbd_dev`
+    do_rbd unmap $rbd_dev || exit $OCF_ERR_GENERIC
+
+    # After the resource has been stopped, check whether it shut down
+    # correctly. If the resource stops asynchronously, the agent may
+    # spin on the monitor function here -- if the resource does not
+    # shut down within the defined timeout, the cluster manager will
+    # consider the stop action failed
+    while rbd_monitor; do
+        ocf_log debug "Resource has not stopped yet, waiting"
+        sleep 1
+    done
+
+    # only return $OCF_SUCCESS if _everything_ succeeded as expected
+    return $OCF_SUCCESS
+
+}
+
+# Make sure meta-data and usage always succeed
+case $__OCF_ACTION in
+meta-data)      rbd_meta_data
+                exit $OCF_SUCCESS
+                ;;
+usage|help)     rbd_usage
+                exit $OCF_SUCCESS
+                ;;
+esac
+
+# Anything other than meta-data and usage must pass validation
+rbd_validate_all || exit $?
+
+# Translate each action into the appropriate function call
+case $__OCF_ACTION in
+start)          rbd_start;;
+stop)           rbd_stop;;
+status|monitor) rbd_monitor;;
+validate-all)   ;;
+*)              rbd_usage
+                exit $OCF_ERR_UNIMPLEMENTED
+                ;;
+esac
+rc=$?
+
+# The resource agent may optionally log a debug message
+ocf_log debug "${OCF_RESOURCE_INSTANCE} $__OCF_ACTION returned $rc"
+exit $rc

=== modified file 'revision'
--- revision	2013-02-13 23:44:37 +0000
+++ revision	2013-03-02 03:56:20 +0000
@@ -1,1 +1,1 @@
-61
+66