2010/2/25 René Nussbaumer <[email protected]>:
> --- /dev/null
> +++ b/tools/cluster-merge
> +LOG_CLUSTER_MERGE = constants.LOG_DIR + "/cluster-merge.log"
This is not used.
> +PAUSE_PERIOD_OPT = cli.cli_option("-p", "--watcher-pause-period",
> default=1800,
> + action="store", type="int",
> + dest="pause_period",
> + help=("Period of time watcher should be"
s/Period of time/Amount of time in seconds/
> + " suspended from running"))
> +class MergerData(object):
> + """Container class to hold data used for merger.
> +
> + """
> + #W0621: Redefining name 'config' from outer scope
> + def __init__(self, cluster, key, instances, nodes, # pylint:
> disable-msg=W0621
Don't redefine names unless there's a really, really good reason for
it (like interface compatibility with a third-party library).
> +class Merger(object):
> + """Handling the merge.
> +
> + """
> + def __init__(self, clusters, pause_period=1800):
Why make pause_period optional?
> + self.work_dir = tempfile.mkdtemp(suffix="cluster-merger")
> + self.ssh_runner = ssh.SshRunner("faked-cluster-merger-cluster")
Shouldn't this use the local cluster name?
> + def _StopMergingInstances(self):
> + """Stop instances on merging clusters.
> +
> + """
> + for cluster in self.clusters:
> + result = self._RunCmd(cluster, "gnt-instance shutdown --all"
> + " --force-multiple")
> +
> + if result.failed:
> + raise errors.RemoteError("Unable to stop instances on %s. "
> + "Fail reason: %s; output: %s" %
> + (cluster, result.fail_reason,
> result.output))
When wrapping strings, whitespace must be on the following line (see
http://code.google.com/p/ganeti/wiki/StyleGuide). Here and in at least
one other place.
> + def _DisableWatcher(self):
> + """Disable watch on all merging clusters, including ourself.
> +
> + """
> + for cluster in ["localhost"] + self.clusters:
> + result = self._RunCmd(cluster, "gnt-cluster watcher pause %d" %
> + self.pause_period)
> +
> + if result.failed:
> + raise errors.RemoteError("Unable to pause watcher on %s. "
> + "Fail reason: %s; output: %s" %
> + (cluster, result.fail_reason,
> result.output))
See above.
The code will soon be ready!
Regards,
Michael
