servicegroup的mc实现

前面的博文分析过servicegroup的实现分为db和mc两种，前面已经说过db了，下来看看mc的实现，这里的mc
全称应该是memcached_client的缩写

class MemcachedDriver(base.Driver):

    def __init__(self, *args, **kwargs):
#首先在init函数中为mc赋值
        self.mc = cache_utils.get_memcached_client(
                expiration_time=CONF.service_down_time)

    def join(self, member_id, group_id, service=None):
        """Join the given service with its group."""

        LOG.debug('Memcached_Driver: join new ServiceGroup member '
                  '%(member_id)s to the %(group_id)s group, '
                  'service = %(service)s',
                  {'member_id': member_id,
                   'group_id': group_id,
                   'service': service})
        if service is None:
            raise RuntimeError(_('service is a mandatory argument for '
                                 'Memcached based ServiceGroup driver'))
        report_interval = service.report_interval
#在join函数中和db的实现一样，都是在要监视的servie中创建一个time，且timer的回调函数为_report_state
#这个join函数是否可以放到
#基类base.Driver 中去实现呢？
        if report_interval:
            service.tg.add_timer(report_interval, self._report_state,
                                 api.INITIAL_REPORTING_DELAY, service)

    def is_up(self, service_ref):
        """Moved from nova.utils
        Check whether a service is up based on last heartbeat.
        """
#按存入mc的规则组装key
        key = "%(topic)s:%(host)s" % service_ref
#判断这个service是否还活着的方法就是判断mc.ge中是否包含key
        is_up = self.mc.get(str(key)) is not None
        if not is_up:
            LOG.debug('Seems service %s is down', key)

        return is_up

    def _report_state(self, service):
        """Update the state of this service in the datastore."""
        try:
            key = "%(topic)s:%(host)s" % service.service_ref

#和is_up函数中相同的规则来组装key，通过mc.set函数保存key和当前的ut时间
            self.mc.set(str(key),
                        timeutils.utcnow())

mc机制和db机制实现基本是一样的，不同在写入的存储介质不一样，db是写入到数据库中.
而mc是写入到mem cache中.